Knowledge Base

Knowledge base

Exported on 01/08/2024
 Knowledge base – Knowledge Base

Table of Contents
1 Inertial Measurements Units ............................................................................. 10
1.1 MEMS Technology ....................................................................................................................11
1.1.1 What is MEMS Technology? ................................................................................................................................... 11
1.1.2 Vibration Handling ................................................................................................................................................. 12
1.2 Accelerometers ........................................................................................................................12
1.2.1 Definition ................................................................................................................................................................ 12
1.2.2 How do we use them.............................................................................................................................................. 12
1.2.3 Principle of operation ............................................................................................................................................ 13
1.2.4 MEMS capacitive technology................................................................................................................................. 13
1.3 Gyroscopes ...............................................................................................................................15
1.3.1 Definition ................................................................................................................................................................ 15
1.3.2 How do we use gyroscopes ................................................................................................................................... 15
1.3.3 Principle.................................................................................................................................................................. 16
1.3.3.1 Coriolis effect ......................................................................................................................................................... 16
1.3.3.2 Sagnac Effect.......................................................................................................................................................... 16
1.4 Magnetometers ........................................................................................................................17
1.4.1 Definition ................................................................................................................................................................ 17
1.4.2 Principle.................................................................................................................................................................. 17
1.4.3 Need of a magnetic calibration ............................................................................................................................. 18
1.5 Factory Calibration and Tests .................................................................................................18
1.5.1 Facility .................................................................................................................................................................... 18
1.5.2 Sensor calibration.................................................................................................................................................. 19
1.5.3 Sensor validation ................................................................................................................................................... 19
1.5.4 Calibration validity................................................................................................................................................. 20

2 Integrated Motion & Navigation sensors .......................................................... 21

2.1 Vertical Reference Unit (VRU/MRU).........................................................................................21
2.1.1 Principle.................................................................................................................................................................. 21
2.2 Attitude & Heading Reference System (AHRS) .......................................................................22
2.2.1 Principle.................................................................................................................................................................. 22
2.3 Inertial Navigation System (INS) .............................................................................................22
2.3.1 Principle.................................................................................................................................................................. 22
2.3.2 GNSS aided INS ...................................................................................................................................................... 23

– 2
 Knowledge base – Knowledge Base

2.3.3 Loosely vs Tightly Coupled INS ............................................................................................................................. 24

2.3.3.1 Loosely coupled integration.................................................................................................................................. 24
2.3.3.2 Tightly coupled integration ................................................................................................................................... 25
2.4 Ship Motion Measurements.....................................................................................................25
2.4.1 Real-Time Ship Motion .......................................................................................................................................... 25
2.4.1.1 Ship motion reference frame ................................................................................................................................ 26
2.4.1.2 Heave step response.............................................................................................................................................. 26
2.4.1.3 Swell Mode ............................................................................................................................................................. 27
2.4.2 Delayed Heave........................................................................................................................................................ 27
2.4.3 Center of Rotation and Deported Heave Operation............................................................................................. 28
2.4.4 Heave Enhanced Altitude ...................................................................................................................................... 29
2.4.5 Post-processed Heave ........................................................................................................................................... 29
2.5 GNSS .........................................................................................................................................29
2.5.1 Definition ................................................................................................................................................................ 29
2.5.2 Ground segment .................................................................................................................................................... 30
2.5.3 Standalone GNSS accuracy ................................................................................................................................... 30
2.5.4 Satellite based Augmentation systems ................................................................................................................ 30
2.5.5 Ground based Augmentation Systems / DGPS..................................................................................................... 30
2.5.6 Real Time Kinematic (RTK) .................................................................................................................................... 31
2.5.7 GNSS as INS aiding................................................................................................................................................. 31
2.6 Real time vs Post-Processed Operation..................................................................................31
2.6.1 Real time operation ............................................................................................................................................... 31
2.6.2 Forward Backward Merge Processing................................................................................................................... 31
2.6.3 Removes the need of having a base station + communication link .................................................................... 32

3 Underlying Maths & conventions ...................................................................... 33

3.1 Reference coordinate frames ..................................................................................................33
3.1.1 Geodetic reference system .................................................................................................................................... 33
3.1.2 Type of coordinates ............................................................................................................................................... 34
3.1.2.1 Cartesian coordinates............................................................................................................................................ 34
3.1.2.2 Geographic (or geodetic) coordinates .................................................................................................................. 35
3.1.2.3 Projection (or plane coordinates) ......................................................................................................................... 35
3.1.3 Notion of altitude................................................................................................................................................... 36
3.1.4 Transformations between different geodetic reference systems and different types of coordinates.............. 37
3.2 Vehicle & body coordinate frame ............................................................................................38

– 3
 Knowledge base – Knowledge Base

3.2.1 Vehicle coordinate frame....................................................................................................................................... 38

3.2.2 Sensor (body) coordinate frame ........................................................................................................................... 38
3.2.2.1 Origin of measurements ........................................................................................................................................ 39
3.3 Orientation / Rotations representation ..................................................................................39
3.3.1 Positive rotation direction..................................................................................................................................... 39
3.3.2 Euler angles ............................................................................................................................................................ 39
3.3.3 Quaternions............................................................................................................................................................ 39
3.3.4 Rotation matrix (Direction Cosine Matrix) ............................................................................................................ 40
3.3.5 Other useful conversion formulas......................................................................................................................... 40
3.3.5.1 Quaternions to DCM............................................................................................................................................... 40
3.3.5.2 Quaternions to Euler.............................................................................................................................................. 41
3.3.5.3 DCM to Euler ........................................................................................................................................................... 41
3.4 Automotive conventions .........................................................................................................41
3.4.1 INS alignment......................................................................................................................................................... 41
3.4.2 Slip angle ................................................................................................................................................................ 41
3.4.2.1 Accuracy ................................................................................................................................................................. 42
3.4.3 Curvature radius .................................................................................................................................................... 43

4 Inertial Sensors Installation .............................................................................. 45

4.1 Inertial sensor installation.......................................................................................................45
4.1.1 Standard Installation ............................................................................................................................................. 45
4.1.2 High Vibrations Considerations............................................................................................................................. 46
4.1.3 Magnetic Consideration......................................................................................................................................... 47
4.2 Accounting for misalignment ..................................................................................................47
4.2.1 Axis misalignment .................................................................................................................................................. 47
4.2.1.1 Axis Example alignment......................................................................................................................................... 48
4.2.2 Fine misalignment ................................................................................................................................................. 49
4.2.2.1 Misalignment residuals Measurments: ................................................................................................................. 50
4.2.3 Vehicle roll, pitch and yaw misalignment............................................................................................................. 51
4.3 GNSS antenna installation.......................................................................................................52
4.3.1 Dual antenna systems considerations.................................................................................................................. 52
4.3.2 GNSS lever arm measurement .............................................................................................................................. 53
4.4 Accounting for Lever arms.......................................................................................................54
4.5 Odometer installation..............................................................................................................54
4.5.1 Mechanical Installation.......................................................................................................................................... 55

– 4
 Knowledge base – Knowledge Base

4.5.2 Electrical installation ............................................................................................................................................. 55

4.5.3 Quadrature vs. Direction odometers .................................................................................................................... 55
4.5.3.1 Quadrature Odometer ........................................................................................................................................... 55
4.5.3.2 Odometer with direction signal ............................................................................................................................ 56
4.6 DVL installation ........................................................................................................................56
4.6.1 Mechanical installation.......................................................................................................................................... 56
4.6.2 Electrical connections ........................................................................................................................................... 57
4.6.2.1 Data output connection......................................................................................................................................... 57
4.6.2.2 Synchronization between DVL and SBG INS......................................................................................................... 57
4.7 Lever arms & alignment calibration........................................................................................58
4.7.1 Real-Time ............................................................................................................................................................... 58
4.7.2 Post-Processing...................................................................................................................................................... 60
4.8 Magnetic calibration ................................................................................................................61
4.8.1 Overview and background..................................................................................................................................... 61
4.8.1.1 The most advanced calibration procedure available on the market .................................................................. 62
4.8.1.2 What can be compensated for and what cannot ................................................................................................. 62
4.8.2 Procedure ............................................................................................................................................................... 62
4.8.2.1 3D Calibration......................................................................................................................................................... 63
4.8.2.2 2D Calibration......................................................................................................................................................... 63
4.8.3 Calibration methods .............................................................................................................................................. 63
4.8.3.1 sbgCenter method ................................................................................................................................................. 63
4.8.3.2 On-board calibration ............................................................................................................................................. 68
4.8.4 Advices and limitations.......................................................................................................................................... 69
4.8.4.1 Cover a maximum of different orientations ......................................................................................................... 69
4.8.4.2 Limited degrees of freedom .................................................................................................................................. 69
4.8.4.3 Keep away from any external sources of distortion............................................................................................. 70
4.8.5 Specific applications.............................................................................................................................................. 71
4.8.6 Magnetic calibration for airborne applications.................................................................................................... 71
4.8.6.1 Light UAV calibration ............................................................................................................................................. 71
4.8.6.2 Airplanes, Helicopter and large UAV applications................................................................................................ 71
4.8.7 Magnetic calibration in marine applications........................................................................................................ 74
4.8.7.1 Light Ship calibration (ASV, ROV, AUV).................................................................................................................. 74
4.8.7.2 Large ship calibration ............................................................................................................................................ 74
4.9 Specific features regarding the application ...........................................................................75
4.9.1 Airborne .................................................................................................................................................................. 75

– 5
 Knowledge base – Knowledge Base

4.9.1.1 Mechanical installation.......................................................................................................................................... 75

4.9.1.2 GNSS setup considerations ................................................................................................................................... 76
4.9.2 Land ........................................................................................................................................................................ 78
4.9.2.1 Mechanical installation.......................................................................................................................................... 78
4.9.2.2 GNSS setup considerations ................................................................................................................................... 80
4.9.3 Marine ..................................................................................................................................................................... 82
4.9.3.1 Mechanical installation.......................................................................................................................................... 82
4.9.3.2 GNSS setup considerations ................................................................................................................................... 83

5 Inertial Sensors Operation................................................................................. 86

5.1 SBG EKF modes of Operation ..................................................................................................86
5.1.1 sbgEkf Introduction ............................................................................................................................................... 86
5.1.2 Basic principle ........................................................................................................................................................ 86
5.1.3 Mode of Operation ................................................................................................................................................. 87
5.1.3.1 Uninitialized mode................................................................................................................................................. 87
5.1.3.2 Vertical Gyro mode................................................................................................................................................. 87
5.1.3.3 Heading Rough alignment procedures ................................................................................................................. 87
5.1.3.4 AHRS mode............................................................................................................................................................. 88
5.1.3.5 Position and velocity initialization........................................................................................................................ 88
5.1.3.6 Full Navigation mode............................................................................................................................................. 88
5.1.4 Heading Observability ........................................................................................................................................... 89
5.1.4.1 Using Single Antenna GNSS................................................................................................................................... 89
5.1.4.2 Using Dual Antenna GNSS ..................................................................................................................................... 89
5.1.4.3 Automotive applications ....................................................................................................................................... 89
5.1.4.4 Using Magnetic Heading ........................................................................................................................................ 89
5.1.5 Improving Navigation Performance...................................................................................................................... 90
5.1.5.1 Lateral velocity constraints ................................................................................................................................... 90
5.1.5.2 Odometer Aiding .................................................................................................................................................... 90
5.1.5.3 Doppler Velocity Log Aiding................................................................................................................................... 90
5.1.5.4 ZUPT Mode ............................................................................................................................................................. 90
5.2 Real time operation with cm level accuracy...........................................................................91
5.2.1 RTK.......................................................................................................................................................................... 91
5.2.1.1 Typical RTK setup................................................................................................................................................... 92
5.2.1.2 RTK corrections formats ........................................................................................................................................ 92
5.2.1.3 Installing a base station......................................................................................................................................... 92
5.2.2 Network RTK (NRTK) .............................................................................................................................................. 92

– 6
 Knowledge base – Knowledge Base

5.2.2.1 NTRIP ...................................................................................................................................................................... 93

5.2.2.2 VBS / VRS ................................................................................................................................................................ 93
5.2.3 PPP.......................................................................................................................................................................... 94
5.3 Ultimate accuracy with post-processing ................................................................................94
5.3.1 Tight Coupling PPK ................................................................................................................................................ 94
5.3.1.1 What is needed....................................................................................................................................................... 95
5.3.1.2 PPK in Qinertia ....................................................................................................................................................... 95
5.3.2 Post-processed PPP ............................................................................................................................................... 97
5.4 Time, synchronization and event management.....................................................................97
5.4.1 INS output latency ................................................................................................................................................. 98
5.4.2 Events ..................................................................................................................................................................... 98
5.4.2.1 Event inputs............................................................................................................................................................ 98
5.4.2.2 Event output........................................................................................................................................................... 99
5.4.3 Synchronizing with external devices .................................................................................................................... 99
5.4.3.1 Network Time Protocol (NTP) ............................................................................................................................. 100
5.4.3.2 Precision Time Protocol (PTP)............................................................................................................................. 100

6 Technology insights ......................................................................................... 101

6.1 INS benefits over MRU in Hydrography.................................................................................101
6.1.1 Principle................................................................................................................................................................ 101
6.1.1.1 MRU Outputs ........................................................................................................................................................ 101
6.1.1.2 INS Outputs .......................................................................................................................................................... 102
6.1.2 INS Measurement's improvement ...................................................................................................................... 102
6.2 How to compare IMU .............................................................................................................103
6.2.1 Full Scale .............................................................................................................................................................. 103
6.2.2 Scale Factor Stability ........................................................................................................................................... 104
6.2.3 Non-linearity ........................................................................................................................................................ 105
6.2.4 One Year Bias Stability......................................................................................................................................... 106
6.2.5 Velocity/Angular Random Walk........................................................................................................................... 107
6.2.6 In Run Bias Instability .......................................................................................................................................... 108
6.2.7 Allan variance ....................................................................................................................................................... 109
6.2.8 Vibration Rectification Error................................................................................................................................ 110
6.2.9 Sampling Rate ...................................................................................................................................................... 111
6.2.10 Bandwidth ............................................................................................................................................................ 111
6.2.11 Resolution ............................................................................................................................................................ 112

– 7
 Knowledge base – Knowledge Base

6.2.12 Orthogonality ....................................................................................................................................................... 113

6.2.13 Cross-Axis Sensitivity ........................................................................................................................................... 113
6.3 Antenna Characteristics: How to choose the correct one....................................................113
6.3.1 Connection: .......................................................................................................................................................... 114
6.3.1.1 Connector type:.................................................................................................................................................... 114
6.3.1.2 Input Impedance:................................................................................................................................................. 114
6.3.2 Power supply:....................................................................................................................................................... 114
6.3.2.1 Active vs Passive:.................................................................................................................................................. 114
6.3.2.2 Voltage:................................................................................................................................................................. 114
6.3.2.3 Current draw: ....................................................................................................................................................... 115
6.3.3 Signal: ................................................................................................................................................................... 115
6.3.3.1 Amplifier gain : ..................................................................................................................................................... 115
6.3.3.2 Constellations :..................................................................................................................................................... 115
6.3.3.3 Frequency band: .................................................................................................................................................. 115
6.4 UART Baud Rate and Output Rate.........................................................................................115
6.4.1 Definition .............................................................................................................................................................. 116
6.4.1.1 Bits and Bytes....................................................................................................................................................... 116
6.4.1.2 UART frame........................................................................................................................................................... 116
6.4.1.3 Baud...................................................................................................................................................................... 116
6.4.1.4 Baud rate .............................................................................................................................................................. 116
6.4.1.5 Output rate ........................................................................................................................................................... 116
6.4.2 Selecting the correct baud rate for a specific amount of data output per second........................................... 117
6.4.3 Calculating the amount of bytes per second that can be output for a specific baud rate ............................... 118
6.4.4 Length of cable vs baud rate ............................................................................................................................... 118
6.4.5 How to check for port saturation ........................................................................................................................ 118

7 sitemap............................................................................................................. 121

– 8
 Knowledge base – Knowledge Base

Learn more about Inertial Sensor technology, Motion, Positioning, Navigation and also Post Processing.

– 9
 Knowledge base – Knowledge Base

1 Inertial Measurements Units

An Inertial Measurement Unit (IMU) is at the core of all inertial systems.
The IMU degrees of freedom typically reflects the number of sensors axes :
• 3 DOF for a gyro triad
• 6 DOF for a full IMU with 3 accelerometers and 3 gyroscopes
IMUs including magnetometers sometime are advertised as 9 degrees of freedom, by counting the magnetic
measurements as well. However, in a practical situation, the magnetic field is directly linked to the device's
orientation, so 6 DoF still stand for an IMU with magnetometers.
For optimal performance, a temperature sensor is also required if a temperature calibration is applied in the IMU.

In this section, you will learn more details about these sensors and the MEMS technology.

• MEMS Technology(see page 11)

• Accelerometers(see page 12)
• Gyroscopes(see page 15)

Inertial Measurements Units – 10

 Knowledge base – Knowledge Base

• Magnetometers(see page 17)

• Factory Calibration and Tests(see page 18)

1.1 MEMS Technology

1.1.1 What is MEMS Technology?

Micro-Electro-Mechanical Systems, or MEMS is a technology that can be defined as miniaturized mechanical and
electro-mechanical elements at waffer scale. This allows MEMSs technology to be embedded on chips that can be
integrated on a printed circuit, like any other electronics component.
Devices are composed of electromechanical elements (moving or not) under the control of integrated
microelectronics.
Functional elements of MEMS are miniaturized structures, sensors, actuators and microelectronics. Microsensors
and microactuators are categorized as transducers which are devices converting a measured mechanical signal in
electrical one.

Micro sensors can be used now for various measurements such as temperature, pressures, magnetic fields or
inertial measurements.
High quality MEMS components have been selected for our IMUs. This MEMS technology provides many advantages
over competing technologies such as mechanical or FOG gyroscopes, or servo accelerometers:
• A miniature design provides smaller, lighter products, enabling new applications to be covered.
• This technology is very robust and provides much higher shock resistance as well as maintenance free
operation.
• MEMS designs provide cost effective solutions compared to other technologies such as FOG or RLG.

Inertial Measurements Units – 11

 Knowledge base – Knowledge Base

MEMS Ring of a gyroscope - Courtesy of Silicon Sensing

MEMS perfomance
Although the same MEMS technology is used for consumer applications such as smart-phones and tablets,
there is a very large performance gap between low cost MEMS and navigation grade MEMS sensors.
SBG Systems only selects high performance sensors only.

1.1.2 Vibration Handling

SBG Systems IMUs has been designed for harsh environments. Specific developments led to efficient vibration
handling.
When exposed to vibrations, an accelerometer or gyroscope will have some increased bias. This vibration effect on
accelerometer is called VRE. So a good starting point is to choose sensors that have low VRE in order to sustain
higher levels of vibrations.
The second point is to design efficient hardware and software signal processing, particularly high frequency
sampling (>1kHz) will prevent aliasing issues and improve sensor data resolution.

1.2 Accelerometers

1.2.1 Definition
An accelerometer is a sensor that measures the specific force (or the rate of velocity change). The specific force
represents all inertial accelerations, including changes of velocity in the local coordinate frame, and reaction to
gravity.
Its output is typically expressed in m/s².

1.2.2 How do we use them

Accelerometers can be used in several ways in inertial systems.

Inertial Measurements Units – 12

 Knowledge base – Knowledge Base

In typical AHRS or Vertical Gyroscopes systems, the accelerometers are used to measure the direction of gravity
vector, that is used as vertical reference.
In modern Inertial navigation Systems, the accelerometer signal is integrated once to obtain velocity information,
and twice to obtain a position information.

Note that accelerometer signal must be used in conjunction with gyroscope information to account for
any rotations.

1.2.3 Principle of operation

An accelerometer can be schematized as a proof mass attached to a spring. The displacement is measured to
estimate the acceleration.

1.2.4 MEMS capacitive technology

Typical MEMS accelerometers employ a capacitive design. A mobile comb is fixed to a fixed comb through a spring.
When no acceleration is present, the mobile comb is centered in such ways that all capacitors have the same value.
However, in presence of accelerations in the sensitive axis, the mobile comb moves away which creates a different
capacitance on each side of the comb. The difference in capacitance is proportional to the acceleration.
A set of analog components and Analog to Digital Converters are used to convert this differential capacitance into
an acceleration in natural units.

Inertial Measurements Units – 13

Knowledge base – Knowledge Base

Inertial Measurements Units – 14

 Knowledge base – Knowledge Base

1.3 Gyroscopes

1.3.1 Definition
A gyroscope is a sensor that measures the rotation rate with respect to an inertial frame.
It will thus measure all the subsequent changes in orientation when it's rotated.
Since the Earth coordinate frame is rotating, a gyroscope lying on the ground will also measure the Earth rotation
rate (approximately 15°/h).
Gyroscope output is typically expressed in °/s, or in rad/s.

1.3.2 How do we use gyroscopes

Rotation rates can be integrated at high frequency to measure precisely the sensor orientation. This principle is
used at the base of all inertial sensors (Vertical Gyroscopes, AHRS, MRU, INS).
They are also used in navigation (Position and Velocity) determination in Inertial navigation Systems (INS).
Highest performance gyroscopes (typically closed loop FOG based) are so accurate that they can measure the Earth
rotation without any bias estimation, and determine a heading without any external heading reference. These
sensors are called Gyrocompasses. There are many other techniques that allow to measure heading without doing
pure gyro-compassing, and this allows to release the constraints on gyro technology.

Inertial Measurements Units – 15

 Knowledge base – Knowledge Base

1.3.3 Principle
SBG Systems gyroscopes use 2 principles to measure rotation:

1.3.3.1 Coriolis effect

Coriolis effect is used by MEMS gyroscopes. When a proof mass is vibrating in the X direction, a rotation on the
perpendicular axis Z, will generate a force causing a corresponding motion on the perpendicular axis Y.
There are several MEMS architectures, typically based on vibrating beams, or vibrating rings.
The actual gyroscope performance may vary depending on MEMS element size and quality, sensor packaging,
mechanical and signal processing.

1.3.3.2 Sagnac Effect

This effect is used by optical gyroscopes such as Fiber Optic Gyroscopes (FOG).
Two light sources are looped traveling in opposite direction: one clockwise, the other one counterclockwise. When
there is no rotation, the light from both sources arrive at the same time, however if there is a rotation one of the
path will be longer and the other-one shorter. This will create interference that can be observed to deduct the
angular speed.

Inertial Measurements Units – 16

 Knowledge base – Knowledge Base

1.4 Magnetometers

1.4.1 Definition
Magnetic sensors are used to measure the Earth magnetic field. These sensors enable a compassing capability by
providing direction to the magnetic North.
Usually magnetic field is expressed in nano Teslas or Gauss. However, compass applications do not really require
specific units since only the direction of magnetic field matters. Thus, a normalized vector (with a norm equal to
1.0) can perfectly describe the direction to Magnetic North.

1.4.2 Principle
Ellipse magnetometers are Anisotropic Magneto Resistance (AMR) based sensors. These are constituted of a tiny
Silicon board and layer of ferromagnetic alloy (see picture below). Resistance depends on the angle between the
current direction and magnetization direction, which is influenced by magnetic field.

The advantages of AMR magnetometers are good sensitivity, and good bias stability.

Inertial Measurements Units – 17

 Knowledge base – Knowledge Base

1.4.3 Need of a magnetic calibration

The Earth magnetic field is particularly weak (typically in the range of 250 to 650 mGauss).
In addition, the materials located in the vicinity may greatly affect the local magnetic field. In particular hard irons
(magnets) can create local deviations of the magnetic field, orders of magnitude higher than Earth magnetic field.
Soft Iron are ferromagnetic objects that react to their local magnetic field by changing its direction and can also
disturb the local magnetic field.
In order to distinguish the weak Earth magnetic field, from the local disturbances, it becomes necessary to perform
an in-situ calibration to map correctly all disturbances and compensate them.
For more information on Ellipse magnetometers calibration, please read the following document: Magnetic
calibration(see page 61)

Periodic Calibration?
Unlike typical low cost magnetic sensors which require regular re-calibration, SBG Systems select only
sensors that don't get magnetized over time and thus, don't require periodic calibration.

1.5 Factory Calibration and Tests

In order to provide best quality sensors, SBG Systems has developed a unique calibration and test procedures.
Highly automated, these procedures are scalable and avoid all the common human errors and make sure that every
sensor delivered meets the specifications.

1.5.1 Facility
Our facility is equipped with multi-axes rotary tables and temperature chambers, capable of simulating all the
environments that your sensor could meet during operation. These tools are powered by an advanvced calibration
software that automates all the process, including environmental and motion simulation, calibration parameters
computation, sensor validation, and report generation.

Inertial Measurements Units – 18

 Knowledge base – Knowledge Base

1.5.2 Sensor calibration

When dealing with sensors error parameters, we consider that a good calibration is always better and more reliable
than on-line sensor estimation. We calibrate and test each product in our factory in order to provide efficient and
defect free units.
The calibration procedure provides:
• Functional and accuracy test of all sensors, and subsystems over full temperature range.
• Gain and bias compensation over full temperature range for accelerometers and gyroscopes,
• Non linearity compensation for accelerometers and gyroscopes over full measurement range,
• Cross-axis and misalignment effects compensation for accelerometers and gyroscopes,
• Gyro-G effect compensation for gyroscopes.
A calibration (and following tests) sequence typically takes 1 to 3 days in order to achieve all the tasks necessary.
This time enables optimal performance of your sensor.

1.5.3 Sensor validation

Once the calibration is performed, all sensors follow a validation step, where the actual calibration parameters are
burned into the sensor's non volatile memory. Real motion is applied on the sensor and measurements are
compared to the reference to ensure that the sensor behavior is perfectly consistent with specifications.
A report automatically generated ensures that only compliant units are delivered.
Following extracts show typical examples of tests performed and how results are presented on the validation
report.

Inertial Measurements Units – 19

 Knowledge base – Knowledge Base

1.5.4 Calibration validity

The validation report guarantees full product specifications for three years following the IMU calibration date. If
your quality policy requests it, SBG Systems can offer a sensor check & calibration service once this period has
expired.
However, SBG Systems has carefully qualified the IMU & INS drift over time, and selected only sensors with an
outstanding stability over time. If used under normal operating conditions, the product should meet its
specifications over life time and should not necessitate regular calibrations.

Inertial Measurements Units – 20

 Knowledge base – Knowledge Base

2 Integrated Motion & Navigation sensors

In this section, you will learn about how integrated navigation and motion sensors are designed, and GNSS
technology.

• Vertical Reference Unit (VRU/MRU)(see page 21)

• Attitude & Heading Reference System (AHRS)(see page 22)
• Inertial Navigation System (INS)(see page 22)
• Ship Motion Measurements(see page 25)
• GNSS(see page 29)
• Real time vs Post-Processed Operation(see page 31)

2.1 Vertical Reference Unit (VRU/MRU)

2.1.1 Principle
A Vertical Reference Unit (VRU) is comprised of an inertial measurement unit (IMU) and filtering to be able to
provide Roll and Pitch. It uses vertical reference (gravity) to stabilize the IMU. The gyroscopes data is fused with the
gravity measurement from the accelerometers in the Kalman filter to provide Roll and Pitch angles.
VRUs take the advantage of gyroscopes to provide accurate roll and pitch during low to medium dynamic
maneuvers. It is also very easy to setup and use. However, the motion precision might degrade in a dynamic
environment due to the impossibility to distinguish accelerations from the gravity (vertical) reference.

A Motion Reference Unit (MRU) is a VRU that can also provide ship motion (Heave, Surge and Sway) in addition to
Roll and Pitch for marine applications. For more information on Heave, follow Ship Motion Measurements(see page
25).

Integrated Motion & Navigation sensors – 21

 Knowledge base – Knowledge Base

2.2 Attitude & Heading Reference System (AHRS)

2.2.1 Principle
An Attitude and Heading Reference System (AHRS) is an inertial sensor comprised of an IMU and another sensor
used for heading determination (typically a magnetometer).
It uses vertical reference (gravity) to stabilize the IMU. The gyroscopes data is fused with the gravity measurement
from the accelerometers in the Kalman filter to provide Roll and Pitch angles. For Heading computation, an aiding
sensor can be used to stabilize the heading such as a magnetometer or a dual antenna GNSS input.

An AHRS takes the advantage of gyros to provide accurate roll, pitch and yaw outputs, in low to medium dynamic
conditions. It is also very easy to setup and use.
However, its overall precision might degrade in a dynamic or magnetically disturbed environment due to the
impossibility to distinguish accelerations from the gravity (vertical) reference.

2.3 Inertial Navigation System (INS)

2.3.1 Principle
An Inertial Navigation System (INS) is a device that relies on a triad of gyroscopes and accelerometers to compute a
6 degrees of freedom navigation output (orientation, velocity and position).

Integrated Motion & Navigation sensors – 22

 Knowledge base – Knowledge Base

The INS is highly efficient for short term navigation. As the sensors are integrated over time, there is a residual error
that tends to increase. Therefore, corrections must be regularly applied on the INS outputs to keep reasonable
errors.

2.3.2 GNSS aided INS

In order to implement these regular corrections, it's very convenient to fuse GNSS receivers with inertial sensor
data: In this scenario, the GNSS short term errors are efficiently corrected by inertial sensors, and inertial drift is
corrected by GNSS. A common way to do this sensor fusion is to use an Extended Kalman Filte (EKF). The major
advantage of EKF is that it tracks the navigation states, but also accounts for sensor biases, and computes an
estimate of uncertainties, giving critical information to user about system performance.
Other external aiding sensors such magnetometer, odometer, or dual antenna GNSS heading sensor can also
complement the system to further improve performance.
The main advantages of INS compared to standalone GNSS, or AHRS systems is to provide high rate orientation and
navigation outputs, as well as a highly accurate solution in all dynamic conditions. An INS is also able to estimate
accurate heading without any dependence on magnetic field.
On the other hand, the INS is a bit more complex to setup as the mechanical installation needs to be known (GNSS
antenna lever arm etc), although it is possible to mitigate this complexity by automated calibration procedures.

Integrated Motion & Navigation sensors – 23

 Knowledge base – Knowledge Base

2.3.3 Loosely vs Tightly Coupled INS

Various schemes of integration exist when coupling inertial sensors and GNSS technology.
The most used ones are the loosely coupled and tightly coupled integrations. We will describe here how they work
and see the pros & cons of each scheme.
In both cases we use an INS to compute short term navigation equations and provide accurate navigation outputs.

2.3.3.1 Loosely coupled integration

In the loosely coupled scheme, the INS is directly fused with GNSS navigation outputs (position, velocity), using
typically an Extended Kalman Fitler.
INS integration errors are corrected by the GNSS position inputs. Accurate and stable navigation outputs are then
delivered by the EKF.
This scheme has a major advantage in terms of simplicity and can be operated on miniature devices, with limited
computational power. It is also the best solution for integration with external receivers such as SAASM (military) or
third party devices using standard NMEA protocol.
However, it's important to understand that the GNSS receiver also needs to run some internal filtering algorithms
(typically a Kalman filter) to compute its navigation solution. When using only position inputs from the receiver, we
will lose some information about the actual GNSS satellites in view. That's why this solution is more dependent on a
good GNSS signal processing.

Integrated Motion & Navigation sensors – 24

 Knowledge base – Knowledge Base

2.3.3.2 Tightly coupled integration

A tightly coupled INS/ GNSS is a more advanced device. In this scheme, the GNSS receiver only computes "RAW
measurements", giving various measures of position with respect to each satellite in view. The GNSS receiver does
not run any navigation filter. All GNSS equations are directly integrated into the INS/GNSS Extended Kalman filter.
This type of integration has the great advantage of giving a much wider ability to identify faulty measurements from
individual satellites, or the possibility to temporarily operate with a limited number of satellites in view (<4). This
provides an excellent robustness to the tightly coupled solution compared to the loosely coupled integration.
In high precision applications (RTK) the tight coupling also allows to increase the availability of centimeter level
accuracy, with faster RTK fix recovery after an outage.
The main limitation of tight coupling is that it is not easily suitable with external GNSS receivers that often do not
provide Raw measurements.

2.4 Ship Motion Measurements

2.4.1 Real-Time Ship Motion

Aside from the EKF, the SBG Systems inertial sensors compute at 50Hz ship motion data from accelerometers
double integration. As this double integration generates drift due to orientation error or sensor bias, the best way to
get a stable output is to use a high pass filter that will remove any constant component in the motion.
The proper behavior of this high pass filter is ensured by an automatic filter tuning, taking into account changes in
sea state conditions. This feature is functional for swell periods up to 15 seconds for real time heave on an Ellipse,
up to 20s on an High Performance products.

Integrated Motion & Navigation sensors – 25

 Knowledge base – Knowledge Base

Due to high pass filter design, the heave, surge and sway data will always return to zero in static conditions.

Surge and Sway validity

Surge and Sway are only relevant in near static applications such as buoys.

2.4.1.1 Ship motion reference frame

The ship motion outputs are in a specific vessel coordinate frame :
• Heave is the vertical position, positive pointing downwards ;
• Surge is the longitudinal position, in the horizontal plane and positive pointing toward the vessel bow ;
• Sway is the transverse position, in the horizontal plane and positive pointing toward the vessel starboard
side.

Surge and Sway availability

Surge and Sway are not available with Ellipse series.

2.4.1.2 Heave step response

If a step is performed, the heave output will show the step and then will smoothly come back to zero. It may take a
few minutes for the output to be stabilized after a step. The following graph shows a typical step response. Note
that the time to recover a 0 heave output may depend on the previous sea state. The overall shape will remain the
same however.

Integrated Motion & Navigation sensors – 26

 Knowledge base – Knowledge Base

2.4.1.3 Swell Mode

On high performance products, the Swell mode is automatically enabled depending on sea conditions to optimize
heave algorithm performance. This mode fuses GNSS information with accelerometer in a smart way to limit phase
and amplitude errors that are inherent to heave algorithms. This mode will be particularly efficient in higher
amplitude swell conditions.
The system determines automatically the best heave computation mode depending on sea state.

2.4.2 Delayed Heave

Available on higher grade units designed for hydrographic applications (Ekinox, Apogee and Navsight Marine for
example), the Delayed Heave algorithm makes use of past measurements to greatly enhance heave performance.
Common phase errors observed in real time Heave operation are seamlessly corrected and the filter will provide
even better performance under long swell period conditions.
The Delayed heave algorithm has a fixed delay of 150s. The output messages have the same format as real time ship
Motion mode and a time-stamp is used to correctly date the ship Motion data.
This algorithm is ideal for applications that don't require strict real time operation such as seabed mapping. The
real time Heave operation remains available to get a first heave estimate before the delayed heave data becomes
available.

Time considerations
As the Delayed heave is a delayed algorithm, the unit must remain turned ON in normal operating
conditions at least 150s before, and 150s after the actual survey path is performed to enable full data
acquisition.

Delayed Surge and Sway?

Integrated Motion & Navigation sensors – 27

 Knowledge base – Knowledge Base

Only vertical ship motion (heave) is available in this Delayed Heave output. Surge and Sway are not
provided in this mode of operation.

2.4.3 Center of Rotation and Deported Heave Operation

When analyzing the heave motion, we can find that part of the heave motion is due to the vessel rotations.
This part differs from one place to another, and is canceled at the Center of Rotations. Another location will affect
the heave in a constant way.
The following picture shows the effect of the rotation induced heave at different locations on the vessel:

In order to optimize the heave performance, the motion sensor placement recommendation may vary depending
on type of vessel or installation.

For smaller survey vessels, placing the sensor directly at the point of interest will ensure best performance (e.g.
direct mounting on top of the MBES sonar) : this location will ensure that no extra noise will be added by calculating
remote heave, and the small Center of Rotation to sensor offset will not create significant errors. This setup is also
the least sensitive to any IMU misalignment or lever arm residuals.
For larger vessels, with a significant Center of Rotation to point of interest lever arm (e.g. above 10m), we
recommend to place the IMU closer to the vessel Center of Rotation in order to minimize constant errors due to long
term roll or pitch offsets.
In that case it's important to setup correctly the IMU misalignments to cancel any constant offsets on the heave
output.
In any case, it is possible to setup the heave output at multiple points of interest within the vessel (e.g. the sonar
head).

IMU misalignment
When the boat is stationary, IMU misalignment with vessel should be precisely accounted for by
mechanical design, or software configuration in order to provide consistent heave values on secondary
points.

Integrated Motion & Navigation sensors – 28

 Knowledge base – Knowledge Base

Center Of Gravity
The Center Of Gravity lever arm in SBG Systems products is not taken into account in the heave
measurements. If required, we recommend to enter this value in the customer survey software used
(Hypack, Qinsy, BeamworX, ...).

2.4.4 Heave Enhanced Altitude

In the marine survey industry, there is often the question whether using the heave output or the Kalman filtered
altitude output is preferred. In case of good GNSS condition the RTK altitude can be very precise and simplifies the
setup as the surveyor don’t need to bother with tide compensation. However, even when fused with inertial
sensors, the RTK altitude might be disturbed in case of difficult GNSS environment such as bridge crossing.
On the other hand, the heave algorithm allows precise relative measurements, without specific errors during
difficult GNSS conditions. However its more complex to use due to the tide compensation needed. The Enhanced
Altitude mode algorithm takes the best of these two worlds by merging the heave output with the RTK altitude,
providing accurate and absolute altitude measurement in both good and challenging GNSS conditions.
This algorithm can only be used with the marine motion profile and in combination with a precise position like RTK
or PPP with fixed carrier ambiguities. This Enhanced Altitude mode can be disabled if needed.

Enhanced Altitude availability

This mode is only available in High performance products designed for hydrographic applications.

2.4.5 Post-processed Heave

In post-processing software such as Qinertia the Heave is re-calculated and improved compared to real time Heave
or delayed Heave thanks to the Forward/Backward and Merged processing.

2.5 GNSS

2.5.1 Definition
GNSS (Global Navigation Satellite System) is a generic word covering all satellite based navigation systems. The first
constellation launched was GPS (Global Positioning System), followed by Russian GLONASS, and more recently
Chinese BEIDOU and European GALILEO. Some other local constellation can complement or replace the system
such as QZSS in Japan, or IRNSS over India.
The major principle used in GNSS is the triangulation : A GNSS receiver is able to measure its distance with respect
to all satellites in view.
Using at least 4 satellites measurement, and knowing the satellites position, the receiver is then able to accurately
compute its own position.

Integrated Motion & Navigation sensors – 29

 Knowledge base – Knowledge Base

2.5.2 Ground segment

In order to get precise satellites position, a "Ground Segment", composed of many ground control stations located
all over the world, compute the satellites position and transfer it back to the satellites.

2.5.3 Standalone GNSS accuracy

The accuracy of a standalone GNSS position ranges form 3 to 10m.

2.5.4 Satellite based Augmentation systems

In order to reduce the total error budget and increase integrity of the GNSS navigation, some augmentation
systems have been developed. The main one is SBAS (satellite based augmentation system), which have several
implementations WAAS, EGNOS, MSAS, GAGAN, and some other are in development.
SBAS accuracy now reaches typically a sub meter level.

2.5.5 Ground based Augmentation Systems / DGPS

The Ground based Augmentation systems uses differential corrections (DGPS or DGNSS) to improve the
performance of a receiver in a local area. It assumes that the major part of the position error budget comes from
atmospheric errors, and can be canceled by using differential computations.

Integrated Motion & Navigation sensors – 30

 Knowledge base – Knowledge Base

DGPS accuracy typically reaches a sub meter level.

2.5.6 Real Time Kinematic (RTK)

RTK is an extension of DGPS using phase and code measurements to achieve centimeter level accuracy.
RTK also takes a great benefit of multi constellations and muti band tracking to increase centimeter level
availability.

2.5.7 GNSS as INS aiding

A GNSS receiver is used as the source of position, velocity and time, but can also be a source of True heading if using
a dual antenna GNSS receiver. These outputs can be in different formats. It will be either ASCII with the standard
NMEA format, or Binary with the format depending on the brand of the GNSS receiver that is being used. RTK GNSS
receivers can be used to deliver centimeter positioning accuracy.
In SBG Systems product variants, all models N and D embed a GNSS receiver. A second external receiver can also be
connected to the high performance series if required. Other INS products accept one or two external GNSS data
inputs to enhance navigation performance.

2.6 Real time vs Post-Processed Operation

2.6.1 Real time operation

Real time navigation systems employ "causal" filters. This mean that the navigation filter can only rely on past and
present measurement to compute navigation outputs.
It is then obvious that during a travel in a tunnel, where the GNSS is not available, the position drift will gradually
increase until we reach the end of the tunnel with maximal error, and then new GNSS position can correct position
drift.

2.6.2 Forward Backward Merge Processing

Now, consider that the processing can be performed after the mission. This is often the case in survey applications.
All the measurements are available at the same time, and it becomes also possible to run the navigation algorithms
starting from the end of the mission, and moving backward in time toward the very first sensor sample.
In the theory of information, the forward pass (similar to real time operation) and backward pass are considered as
independent, because at a given epoch, both passes were computed using a completely different set of
information.
It becomes then possible to "merge" those two independent processing to form a new output, with a better
accuracy than the best of both passes!
The example below shows how the position error evolves when traveling in a tunnel during 60s. We can see that the
error of the position in forward processing grows exponentially until the GNSS is back after 60s. The same can be
seen in the backward data where we lose the GNSS at 60 seconds, and the error grows exponentially until the GNSS
is back at the 0 seconds mark. When we merge and smooth this data our error is decreased significantly, with the
highest error value being in the middle of the GNSS outage.

Integrated Motion & Navigation sensors – 31

 Knowledge base – Knowledge Base

This of course improves greatly performances during GNSS outages, but it also helps with overall performances. It
will also improve heading performancesin low dynamics applications with a single antenna GNSS receiver for
example as the Extended Kalman filter doesn't have to rely as much on predictions since the full history is known
for the whole file.

2.6.3 Removes the need of having a base station + communication link

Another advantage of using post-processing is that it is possible to import a base station as long as the Raw GNSS
data was logged in real-time. This means that in real-time it is not necessary anymore to have a GNSS base station
used, that will save time and money to the user as he won't need to install a base station and a radio link between
the base station and the rover. There is also no need to use a NTRIP link in real-time, so no Internet connection or
computer running an NTRIP client software are needed anymore.
For more Information on RTK in real-time vs PPK in post processing, click on this link: [Real time operation with cm
level accuracy(see page 91)]

Integrated Motion & Navigation sensors – 32

 Knowledge base – Knowledge Base

3 Underlying Maths & conventions

Although this section requires some mathematics skills, it may be important to spend some time to fully
understand how navigation measurements (position, and velocities) and orientations are represented.
You will learn how they are represented, what is a reference frames and which one are used.

Inertial frame
An inertial frame is a frame in which Newton's laws of motion apply. It is therefore not accelerating, but
can be in uniform linear motion. All Inertial sensors (accelerometers, gyroscopes) produce measurements
relative to an inertial frame.

• Reference coordinate frames(see page 33)

• Vehicle & body coordinate frame(see page 38)
• Orientation / Rotations representation(see page 39)
• Automotive conventions(see page 41)

3.1 Reference coordinate frames

This article is a simplified version of the geodesy skills needed for inertial navigation. Should you want further
information, we recommend you visit geodetic sites such as: https://itrf.ign.fr/general.php

3.1.1 Geodetic reference system

A geodetic reference system is a group of elements that enables to locate points using coordinates. Elements are:
• Origin of the coordinates system
• Orientation of axis
• Scale of axis
• Reference ellipsoid
The most important geodetic reference systems are the different realizations of the ITRS (International Terrestrial
Reference System) which are realized by the international scientific community. These realizations are produced at
the same time than technologies progress and the number of data increases to improve results. It is then better to
use the latest realization to get the best accuracy. Currently it is the ITRF2014 (International Terrestrial Reference
Frame 2014).
As the GNSS technology has been created by the US defense, they developed their own geodetic reference system
that corresponds to the different realizations of the WGS84 (World Geodesic System 1984). Initially, all positions
obtained by GNSS measurements were expressed in WGS84. It explains it is the most known one. The latest
realization, G1674, is aligned with the ITRF2014 but some small differences make them agree at the centimeter level
only. For example, the reference ellipsoid of the two systems have 0.1mm variation for the semi-minor axis.

Coordinate reference frame vs Augmented GNSS solution

The solution obtained by GNSS depends on the type of measurements. For RTK or PPK positioning, the solution
is in the base station coordinates system. For PPP positioning or using orbits and clock corrections from IGS,
coordinates are expressed in ITRF2014.

Underlying Maths & conventions – 33

 Knowledge base – Knowledge Base

ITRF2014 vs WG84
As the two geodetic reference systems (ITRF2014 and G1674) are really close, there is no transformation
parameters given and it’s possible to consider that they are the same system (positions obtained by GNSS are
rarely more accurate than one centimeter).

3.1.2 Type of coordinates

For each geodetic reference system, it is possible to use three types of coordinates : Cartesian, geographic or
projected.

3.1.2.1 Cartesian coordinates

It is the easiest way to locate a point in a 3D space. It uses the three orthogonal axis (commonly X, Y, Z) of the
geodetic reference system.

For example, the ITRF2014 has for origin the center of mass of the Earth including oceans and atmosphere. The Z
axis is pointing to the north rotation pole. The XOY plane coincides with the equatorial plane. Orientation of the X
and Y axis are more difficult to understand and have no real interest here. The scale is chosen to correspond to the
meter.

ECEF (Earth Centered Earth Fixed)

If the origin of the coordinate system is close to the Earth center of mass, and the X and Y axis rotate
around the Z axis with the Earth (so it can't be used as an inertial reference frame), this system type is
called ECEF (Earth Centered Earth Fixed) cartesian frame.

Although this type of coordinates is very easy to use, it’s principally utilized by the scientific community. Most users
prefer the geographic coordinates.

Underlying Maths & conventions – 34

 Knowledge base – Knowledge Base

3.1.2.2 Geographic (or geodetic) coordinates

Coordinates above can be also express in geographic form. By using the reference ellipsoid associated to the
geodetic reference system, a point with Cartesian coordinates is projected on the surface of the ellipsoid to
calculate two coordinates : longitude (λ) and latitude (ϕ) which are angles. Then a distance can be determined from
the center of the ellipsoid: ellipsoidal height.
The longitude (λ) is the angle in the equatorial plane from the origin meridian to the projection of the point of
interest onto the equatorial plane.
The latitude (ϕ) is the angle in the meridian plane from the equatorial plane to the ellipsoid normal. Note that, in
most situations, the ellipsoid normal will not intersect with the center of the Earth.
The ellipsoidal height (he or h) is a simple distance between the point and its projection on the ellipsoid.

3.1.2.3 Projection (or plane coordinates)

When working on a small area, it is possible to use a defined projection for a geodetic reference system. This type of
coordinates refers to a 2D space where all the points of the reference ellipsoid are projected on a plane (there are
many ways to do a projection with different effects on the result). The advantage of a such a coordinates system is
its ease of use for map creation. However, as the reference surface is a transposition from an ellipsoid to a plane,
some deformations appear. The commonly used coordinates are E, N (East, North).
It’s important to understand that a projection is associated to a geodetic reference system so two projections don't
always have the same base.

The NED frame (North East Down)

In some cases requiring relative positioning, a local specific Geodetic frame is used in order to ensure best accuracy.
In navigation this local frame refers to the North, East, Down rectangular frame (NED).
This frame is obtained by fitting the local ellipsoid shape in a tangent plane at the current position. This coordinate
frame is attached to a fixed point relatively to the Earth surface.
The X axis is turned toward North, the Z axis turned down, along the local ellipsoid normal, and the Y axis completes
the right hand rule, pointing East.
As it's impossible to perfectly fit the ellipsoidal shape in a plane, this frame is only suitable for local measurements.

Underlying Maths & conventions – 35

 Knowledge base – Knowledge Base

Navigation algorithms Reference frame

The navigation algorithms internally account for this frame rotation when the vehicle moves at high speed
in order to ensure best navigation performance.

3.1.3 Notion of altitude

The altitude (H) is a notion extremely complex, so it's will be briefly developed here. To be defined, it’s necessary to
explain what is the geoid (which can be understood as the Mean Sea Level).
The geoid is an equipotential surface of gravity. It means it is a surface where the norm of each point corresponds
with the local gravity direction. The latest world model is the EGM2008 (Earth Gravity Model). As the local gravity
depends on a lot of things, its surface is absolutely not regular as the ellipsoid can be, so they are not equal.
Altitude is then a distance between a point and its projection on the geoid with distance following the gravity
direction. It's a physical definition.
It is possible to convert ellipsoidal height into altitude by using the gravity model. In the schema below, he is the
ellipsoidal height, H is the altitude and n is the geoid undulation. So it's easy to deduct the relation: he = H + n.

Underlying Maths & conventions – 36

 Knowledge base – Knowledge Base

INS MSL
The INS provides both altitude referenced to the Ellipsoid and Mean Sea Level using a built in EGM96
model with 15 arc sec resolution.

3.1.4 Transformations between different geodetic reference systems and different

types of coordinates
Sometimes, you can have positions expressed in several geodetic reference systems or different types of
coordinates so you need to make them all compatible. For this, the graphic below shows the steps to go from one
system to another. You will notice there are multiple ways to convert coordinates to an other geodetic reference
system but the best to keep accuracy is to use the transformation with seven parameters (three rotations, three
translations and a scale factor). This can only be done with the Cartesian coordinates so you have to start by
transforming your coordinates to their Cartesian form.

14 parameters Datum shifts

If the geodetic reference system takes in consideration the tectonic plates movement, the 7 parameters
transformation becomes a 14 parameters transformation (velocity of all translations, rotations and scale factor
are added).

Underlying Maths & conventions – 37

 Knowledge base – Knowledge Base

3.2 Vehicle & body coordinate frame

3.2.1 Vehicle coordinate frame

Depending on the application, a vehicle coordinate frame is defined as follow : X axis is pointing in Forward
direction, Z axis is pointing Down, and Y axis, thanks to right hand rule is pointing to the right of vehicle.

3.2.2 Sensor (body) coordinate frame

This frame is attached to the IMU.
The following diagram shows the body coordinate frame as configured by default. In most situations, the body
coordinate frame must be aligned with the vehicle coordinate frame. Sensor alignment in the vehicle can be
adjusted by software if the sensor coordinate frame cannot be aligned mechanically. Check section Accounting for
misalignment(see page 47) for more details about this software alignment.

Underlying Maths & conventions – 38

 Knowledge base – Knowledge Base

3.2.2.1 Origin of measurements

We have defined the sensor axes directions, but we also need to know where the Origin of this coordinate frame is.
This coordinate frame origin is the intersection of the three accelerometers and corresponds to the center of
velocity and position measurements.
This origin must be considered when measuring lever arms.

A circular black and white symbol in the mechanical drawings defines and locates this Origin of
measurements.

3.3 Orientation / Rotations representation

3.3.1 Positive rotation direction

According to the “Right Hand Rule”, the positive direction for rotations is clockwise in the axis direction.

3.3.2 Euler angles

Euler angles are a commonly used representation of spatial orientation. Euler angles are in fact a composition of
rotation from the Local Geodetic Coordinates System. This orientation is defined by the sequence of the three
rotations around the Local Frame X, Y and Z axes.
Euler angles are widely used because they are easy to understand. The three parameters : Roll, Pitch and Yaw define
rotations around the fixed frame's axes :
• Roll (φ) : Rotation around X axis defined between [-π ; π] ;
• Pitch (θ) : Rotation around Y axis defined between [-π/2 ; π/2] ;
• Yaw (ψ) : Rotation around Z axis defined between [-π ; π].

Gimbal Lock effect

As Euler angles suffer from a singularity called "Gimbal lock”, when Pitch approaches ± π/2,we do not advise to
use Euler angles if the device has to be used in a wide range of orientations. Quaternions and rotation matrices
do not have any singularity.

3.3.3 Quaternions
Quaternions are an extension of complex numbers as defined here :
$$Q = q_{0} + i \cdot q_{1} + j \cdot q_{2} + k \cdot q_{3}$$
Where i, j, k are imaginary numbers.

Underlying Maths & conventions – 39

 Knowledge base – Knowledge Base

We also can define the complex conjugate of Q :

$$\overline Q = q_{0} - i \cdot q_{1} - j \cdot q_{2} - k \cdot q_{3}$$
And the norm of Q :
$$\| Q \| = \sqrt{Q \cdot \overline Q}$$
Particular quaternions such as |Q| = 1 can represent a complete definition of the 3D orientation without any
singularity.
Quaternions algebra do not require a lot of computational resources, they are therefore very efficient for
orientation representation.
The inverse rotation of Q is defined by its complex conjugate.

3.3.4 Rotation matrix (Direction Cosine Matrix)

The Direction Cosine Matrix (DCM) is a rotation matrix that transforms one coordinate reference frame to another.
Rotation matrices are a complete representation of a 3D orientation, thus there is no singularity in that model.
A DCM locates three unit vectors that define a coordinate frame. Here the DCM transforms the body coordinate
frame to the Local NED coordinates. The DCM is the combination of the three rotation matrices RM(φ), RM(θ) and
RM(ψ) respectively around Local Geodetic (NED) X, Y and Z axes :
$$DCM = RM_{\psi} \cdot RM_{\theta} \cdot RM_{\varphi}$$
$$DCM = \begin{pmatrix} \cos \psi & - \sin \psi & 0 \\ \sin \psi & \cos \psi & 0 \\ 0 & 0 & 1 \end{pmatrix} \cdot
\begin{pmatrix} \cos \theta & 0 & \sin \theta \\ 0 & 1 & 0 \\ - \sin \theta & 1 & \cos \theta \\ \end{pmatrix} \cdot
\begin{pmatrix} 1 & 0 & 0 \\ 0 & \cos \varphi & - \sin \varphi \\ 0 & \sin \varphi & \cos \varphi \end{pmatrix}$$
$$DCM = \begin{pmatrix} \cos \theta \cdot \cos \psi & \sin \varphi \cdot \sin \theta \cdot \cos \psi - \cos \varphi
\cdot \sin \psi & \cos \varphi \cdot \sin \theta \cdot \cos \psi + \sin \varphi \cdot \sin \psi \\ \cos \theta \cdot \sin \psi
& \sin \varphi \cdot \sin \theta \cdot \sin \psi + \cos \varphi \cdot \cos \psi & \cos \varphi \cdot \sin \theta \cdot \sin
\psi - \sin \varphi \cdot \cos \psi \\ \ - \sin \theta & \sin \varphi \cdot \cos \theta & \cos \varphi \cdot \cos \theta
\end{pmatrix}$$
As for any rotation matrix, the inverse rotation equals to the transposed matrix:
$$DCM^{-1} = DCM^{T}$$
In order to transform a vector expressed in the Body coordinate system into the NED frame, user will use the DCM as
expressed below:
$$V_{NED} = DCM \cdot V_{body}$$
Reciprocally:
$$V_{body} = DCM^{T} \cdot V_{NED}$$

3.3.5 Other useful conversion formulas

Some other conversion formulas can be useful for many users, and are listed below:

3.3.5.1 Quaternions to DCM

It may be useful to compute a DCM based on the quaternions parameters :
$$DCM = \begin{pmatrix} 2 \cdot q_{0}^{2} + 2 \cdot q_{1}^{2} - 1 & 2 \cdot q_{1} \cdot q_{2} - 2 \cdot q_{0} \cdot
q_{3} & 2 \cdot q_{0} \cdot q_{2} + 2 \cdot q_{1} \cdot q_{3} \\ 2 \cdot q_{1} \cdot q_{2} + 2 \cdot q_{0} \cdot q_{3} & 2
\cdot q_{0}^{2} + 2 \cdot q_{2}^{2} - 1 & 2 \cdot q_{2} \cdot q_{3} - 2 \cdot q_{0} \cdot q_{1} \\ 2 \cdot q_{1} \cdot
q_{3} - 2 \cdot q_{0} \cdot q_{2} & 2 \cdot q_{2} \cdot q_{3} + 2 \cdot q_{0} \cdot q_{1} & 2 \cdot q_{0}^{2} + 2 \cdot
q_{3}^{2} - 1 \end{pmatrix}$$

Underlying Maths & conventions – 40

 Knowledge base – Knowledge Base

3.3.5.2 Quaternions to Euler

Here is quaternions translated into Euler angles.
$$\varphi = \tan^{-1}{\left( \dfrac{2 \cdot q_{2} \cdot q_{3} + 2 \cdot q_{0} \cdot q_{1}}{2 \cdot q_{0}^{2} + 2 \cdot
q_{3}^{2} - 1}\right)}$$
$$\theta = -\sin^{-1}{\left( 2 \cdot q_{1} \cdot q_{3} - 2 \cdot q_{0} \cdot q_{2}\right)}$$
$$\psi = \tan^{-1}{\left( \dfrac{2 \cdot q_{1} \cdot q_{2} + 2 \cdot q_{0} \cdot q_{3}}{2 \cdot q_{0}^{2} + 2 \cdot q_{1}
^{2} - 1}\right)}$$

3.3.5.3 DCM to Euler

Finally, DCM matrixis converted into Euler angles.
$$\varphi = \tan^{-1}{\left( \dfrac{DCM_{32}}{DCM_{33}}\right)}$$
$$\theta = -\sin^{-1}{\left( DCM_{31} \right)}$$
$$\psi = \tan^{-1}{\left( \dfrac{DCM_{21}}{DCM_{11}}\right)}$$

3.4 Automotive conventions

Automotive applications sometimes involve specific measurements derived from the standard INS output. The
most common outputs are the body velocity, the slip angle and the curvature.

3.4.1 INS alignment

It is important to understand that automotive specific outputs accuracy are affected by the INS intrinsic
performance but also by the mechanical installation within the vehicle. Especially, the INS has to be perfectly
aligned with the vehicle frame to ensure valid and consistent measurements.
If for example, the INS is installed with a 1° heading misalignment, the slip angle measurement will be centered
around 1° instead of 0°.
Moreover, we consider that the nominal vehicle slip angle and curvature measurements are typically computed at
the center of the non steering axle (in general the rear axle).
However it is possible to evaluate these quantities at another location of the vehicle by either installing the INS
where you would like the measurements or by using deported monitoring points.

When using automotive motion profiles, the INS Kalman Filter estimates and refines any pitch and yaw
misalignment to ensure optimal navigation performance. However, this estimated misalignment is only used
internally and thus doesn’t impact the data returned by the product.

3.4.2 Slip angle

The Slip angle is the angle between the course over ground (direction towards which the vehicle is moving) and the
heading (direction where the vehicle longitudinal axis is pointing).

Underlying Maths & conventions – 41

 Knowledge base – Knowledge Base

The following formula is used to compute the slip angle in radians:

$Slip_{angle} = Track_{course} - TrueHeading$
Where Track course is the course over the ground :
$$Track_{course} = tan^{-1} ({V_E \over V_N})$$
Since the course over ground is only defined when the vehicle is moving, this output is only valid in case of velocity
above 1 m/s. In addition, if you see a constant slip angle value, it means the INS is not properly aligned (heading
angle mainly) with the vehicle frame.

vehicle behavior vs Motion profiles

In case the vehicle is expected to experience regular slips/drifts, we recommend the use of motion profiles that
do not use automotive constraints assumptions like the Airplane motion profile for example.

Slip Angle measurement location

The slip angle measurement is using the 2D velocity information and as such is different at each vehicle
location. The most common slip angle measurement is done at the vehicle center of rotation.

3.4.2.1 Accuracy
The accuracy of Slip angle output will be directly impacted by several parameters:
• INS True heading accuracy
• INS velocity accuracy
• Vehicle speed
$$Slip_{acc} = {\sqrt {Heading_{acc}^2 + tan^{-1} {\left({Vel_{acc} \over Speed}\right)^2}}}$$

The accuracy is generally provided for a typical speed of 50 Km/h

Underlying Maths & conventions – 42

 Knowledge base – Knowledge Base

The following table shows example of expected accuracy at different speed for an ELLIPSE-D in RTK mode (with
typical Heading accuracy at 0.1° and velocity at 0.05 m/s):

Speed (km/h) Track Accuracy (°) Slip Accuracy (°)

10 0.29 0.30

20 0.14 0.17

50 0.06 0.11

100 0.03 0.10

200 0.01 0.10

3.4.3 Curvature radius

The radius of curvature reflects instantly how sharp the vehicle is turning. In case of constant speed and constant
angular velocity, like in a roundabout, the curvature radius will be easily compared to the roundabout radius.

The following formula is used to compute the curvature radius in meters:

$$curvature = {\ Vel_{2D} \over w_{Down}}$$
Since the curvature is infinite in case of straight line driving, the value is only computed and valid for angular rate
above 1 °/s and velocity above 1 m/s.

Curvature measurement location

Underlying Maths & conventions – 43

 Knowledge base – Knowledge Base

The curvature measurement is using the 2D velocity information and as such is different at each vehicle
location. The most common curvature measurement is done at the vehicle center of rotation.

Underlying Maths & conventions – 44

 Knowledge base – Knowledge Base

4 Inertial Sensors Installation

SBG Systems provides a range of inertial sensors with some specificity:
Part of the product lines are all fully integrated sensors, with a self contained package including IMU, GNSS and all
computations features, when other solutions have separated IMU and a processing enclosures.
Same installation recommendations apply for all type of products. However, when installing separated IMU
products, processing enclosure can be installed anywhere.

Reference definition
In the following sections, the term “sensor” will always refer to the IMU location.

• Inertial sensor installation(see page 45)

• Accounting for misalignment(see page 47)
• GNSS antenna installation(see page 52)
• Accounting for Lever arms(see page 54)
• Odometer installation(see page 54)
• DVL installation(see page 56)
• Lever arms & alignment calibration(see page 58)
• Magnetic calibration(see page 61)
• Specific features regarding the application(see page 75)

4.1 Inertial sensor installation

4.1.1 Standard Installation

The normal orientation in the vehicle frame is to align sensor X axis to the vehicle forward direction. Sensor Z axis
should be turned down.
When this mechanical alignment is not possible, the IMU misalignment with respect to the vehicle coordinate frame
must be measured, as described in this link Accounting for misalignment(see page 47)

Inertial Sensors Installation – 45

 Knowledge base – Knowledge Base

IMU typical placement in marine applications

As a rule of thumb, SBG Systems sensors can be placed anywhere in the vehicle. However, in case of large vehicles
or vessels, we recommend to place the sensor within 10 meters around the Centre of Rotations.
In any case, the primary lever arm between the sensor and the Centre or Rotation must be measured within 5 cm
accuracy. It is the signed distance FROM the sensor TO the Centre of Rotation, expressed in the vehicle coordinate
frame.
A correct measurement will ensure optimal performance, and particularly in the following applications:
● Marine applications. The heave motion computation is dependent on a good lever arm measurement
● Automotive applications, where the main lever arm is used internally to take into account the motion
constraints assumed in this type of application.

4.1.2 High Vibrations Considerations

SBG Systems has designed IMUs with high quality MEMS sensors combined with high sampling frequency as well as
efficient anti aliasing FIR filters to limit vibration issues as much as possible. Nevertheless, a good mechanical
isolation will ensure getting the full sensor performance:
High amplitude vibrations can cause a bias in accelerometer reading. Thanks to a superior factory calibration, this
effect is limited. Nevertheless it cannot be fully avoided. This effect is called the VRE (Vibration Rectification Error)
and comes from the internal accelerometer non-linearity.
Ultimately, very strong vibrations cause the sensor to saturate. The bias observed will be drastically increased,
leading to a huge error on orientation.

Inertial Sensors Installation – 46

 Knowledge base – Knowledge Base

Note: SBG Systems typically recommends a high accelerometer range to reduce the VRE effect for most application,
except Marine application which need very precise acceleration measurement.
Warning: Heave and delayed heave computations is more sensitive to vibrations than other algorithms. When using
the heave outputs, please take care to reduce as much as possible the vibration level to enable full performance.

4.1.3 Magnetic Consideration

When the internal magnetometer is used as heading reference, care should be taken with ferromagnetic
environment.
Ferromagnetic materials or magnets that are placed in the vicinity of the device can generate error in the
magnetometers readings by distorting the magnetic field. High current power supplies or the associated wires may
also generate magnetic fields.
The sensor should be placed as far as possible from ferromagnetic materials, particularly those who can be moved
independently with respect to the sensor. In practice placing the device more than 2 meter away from disturbing
materials is enough to avoid generating error.
In most cases, a calibration procedure can be performed to map the magnetic distortions and therefore get the full
performance of the unit. The calibration can compensate both Hard and Soft iron interference.

Field Magnetometer Calibration

See Hard & Soft Iron calibration Manual for more information about the magnetometers calibration
procedure on this link Magnetic calibration1

Unexpected Magnetic disturbances

Some disturbances cannot be predicted: a magnet passing suddenly near the device or a cell phone
communication for example.

The internal EKF is able to cope with short term magnetic disturbances. Ultimately if magnetic field direction
changed for a long period, the heading will be realigned to the new magnetic field direction.

When the internal magnetometers are not in use, the magnetic influence on performance is weak but very
strong magnetic fields can affect gyroscopes performance and such high amplitude magnetic fields should
be avoided.

4.2 Accounting for misalignment

The sensor alignment procedure involves two steps: an axis alignment, and a fine alignment. Some aiding sensors
must also take into account misalignment, that will be measured like it has been done for the IMU, comparing the
external sensor with vehicle coordinate frame.

4.2.1 Axis misalignment

The following example shows how to measure IMU axis misalignment. The IMU axes must be compared to the
Vehicle axes as follows:

1 https://support.sbg-systems.com/display/KB/Magnetic+calibration

Inertial Sensors Installation – 47

 Knowledge base – Knowledge Base

IMU AXIS Vehicle direction

X LEFT

Y FRONT

Z DOWN

4.2.1.1 Axis Example alignment

Inertial Sensors Installation – 48

 Knowledge base – Knowledge Base

4.2.2 Fine misalignment

Once axes axis misalignment is performed, the small residual angles must then be measured as follow.
Misalignment angles correspond to the residual rotation required to pass from the IMU coordinate frame to the
vehicle coordinate frame. In our example, alpha corresponds to the mis-heading and its sign is negative.
Most applications will only have low angles on roll and pitch misalignment. If large angles on roll and pitch are
expected (> 5°), user must consider the rotation composition order: roll, then pitch, then yaw.

Mis Angles Value

misroll Not Shown

mispitch Not shown

misheading -α negative)

Inertial Sensors Installation – 49

 Knowledge base – Knowledge Base

4.2.2.1 Misalignment residuals Measurments:

Inertial Sensors Installation – 50

 Knowledge base – Knowledge Base

Once the fine misalignment angles are measured and entered into the device configuration, the sensor coordinate
frame is assumed to be aligned with the vehicle coordinate frame.

4.2.3 Vehicle roll, pitch and yaw misalignment

When the sensor is installed on a vehicle that lies on the ground (not applicable to boat), it is easy to estimate the
roll and pitch misalignment using a simple procedure.
The procedure is applicable on any flat surface (not necessarily horizontal). If we consider a car, it consists in:
• Parking the car in reverse direction on a parking slot, then once stopped, measure the roll and pitch angles.
• Parking the car in forward direction on the same slot. Measure the roll and pitch angles.
• Use the mean roll and mean pitch as respective roll misalignment and pitch misalignment settings in the
configuration page.

Inertial Sensors Installation – 51

 Knowledge base – Knowledge Base

Measuring the INS heading misalignment is however much more complicated. Several methods can be used to do
so as soon as you can guarantee your vehicle is a standard car with non steering rear wheels:
• You should first enter the measured roll and pitch misalignment in the INS configuration.
• Then drive your car in straight lines with sections above 50 km/h and make sure you don’t drift.
• Use the mean slip angle value as a direct measurement of the heading misalignment.
• You can also use Qinertia to automatically estimate and display yaw, pitch and vehicle Center of Rotations
lever arms

4.3 GNSS antenna installation

GNSS antenna must be fixed with respect to the sensor. It should have a clear view of sky.

4.3.1 Dual antenna systems considerations

Dual antenna systems installation requires special care in order to obtain optimal performance:
• Use two antennas of the same type, connected to same cables with identical lengths
• Both antennas must be turned the same way (connectors or North mark oriented in same direction with
respect to the vehicle)
• Both antennas must have a clear view of sky
• In case some masks caused by the vehicle cannot be avoided, it should be minimized, and both
antenna should share the same masks. In other words, it's important that both antenna see the same
satellites.
• Small patch antennas must be placed on a ground plane, and ideally, more than 20cm away from the
ground plane edges
• The Baseline (horizontal distance between two antennas) should be maximized for best performance,
although it is possible to operate on short baseline:
• 30cm will typically lead to sub-degree heading accuracy
• At least one meter is recommended in manned vehicles (as soon as the vehicle size allows this).
• 2m or more is recommended for high precision applications
• Vertical distance between the antennas will not impact the accuracy as long as the horizontal distance is
enough and the lever arms are entered correctly in the product.

Inertial Sensors Installation – 52

 Knowledge base – Knowledge Base

4.3.2 GNSS lever arm measurement

Once installed, both GNSS antenna lever arms must be measured.
The GNSS lever arm is the signed distance, expressed in the vehicle coordinate frame, FROM the sensor's center of
measurement, TO the GNSS antenna phase center. It must be measured within 5cm accuracy.

Inertial Sensors Installation – 53

 Knowledge base – Knowledge Base

4.4 Accounting for Lever arms

All lever arms are considered in the vehicle (body) coordinate frame and are measured FROM the IMU, TO the point
of interest.
Here is an example showing a GNSS antenna lever arm measurement:

4.5 Odometer installation

All our INS models provide an odometer input which can greatly improve performance in challenging environments
such as urban canyons.
The odometer provides a reliable velocity information even during GNSS outages. This increases significantly the
dead reckoning accuracy.
Our products support:
• Quadrature output or compatible odometers with forward and reverse directions.
• CAN vehicle velocity messages (fully configurable) for setup with direct interface with vehicle’s ODBII
connector when using the Ellipse series.

Inertial Sensors Installation – 54

 Knowledge base – Knowledge Base

Odometer integration is made really simple as the EKF will finely adjust odometer's gain and will correct
residual errors in the odometer alignment and lever arm.

4.5.1 Mechanical Installation

The Odometer has to be placed on a non steering wheel (rear wheel in most applications).
The Odometer lever arm has to be measured. It is the signed distance, expressed in the vehicle coordinate frame,
FROM the IMU TO the point of contact between the ground and the tire where the Odometer is installed. It has to be
measured with 5cm accuracy.

4.5.2 Electrical installation

SBG Systems INS devices support several DMI (Distance Measurement Instrument) devices and conventions. You
can connect simple pulse odometer to more complex quadrature wheel encoders that provide both a velocity and a
direction of travel.
For more information, check out the dedicated page(see page 55).

4.5.3 Quadrature vs. Direction odometers

4.5.3.1 Quadrature Odometer

Quadrature odometers are very common and efficient to provide both a velocity as well as the direction of travel.
Two signals are used to report the velocity, the channel A and channel B that is out of phase of 90°.
The convention for a forward (positive) odometer velocity is signal A precede signal B as shown in the diagram
below:

Inertial Sensors Installation – 55

 Knowledge base – Knowledge Base

4.5.3.2 Odometer with direction signal

Some odometer has one channel for the velocity (reported with pulses) and an other signal that provides the
direction of travel as a logic high or low voltage. As you can see below, if the odometer B signal is high, the velocity
is considered to be positive (forward direction).

4.6 DVL installation

4.6.1 Mechanical installation

The DVL must be rigidly fixed to the vessel structure.
It is recommended to align the DVL forward mark toward the vessel bow.
In this case, the nominal alignment angle to enter in the inertial system configuration is +45°.
Note that the native DVL instrument frame (displayed DVLX and DVLY on the diagram) is a left handed frame
whereas the SBG Systems products operate in a right handed coordinate system. Therefore we internally invert the

Inertial Sensors Installation – 56

 Knowledge base – Knowledge Base

DVLX and DVLZ velocity directly on the received sensor input before applying user alignment angles.
Following diagram shows DVL instrument frame as well as the inverted DVL-X (pink arrow) axis that is used for DVL
to vessel misalignment determination.
The lever arm from the IMU to the DVL must also be measured accurately.

4.6.2 Electrical connections

The typical DVL electrical integration takes care of two main aspects:
1. The DVL data output
2. Synchronization between DVL and INS

4.6.2.1 Data output connection

Regarding the data output, a simple serial connection is generally used, by connecting the DVL serial port to one of
the SBG INS available serial ports (eg. PORT A Rx), either in RS-232 or in RS-422.

4.6.2.2 Synchronization between DVL and SBG INS

DVL Master
In this mode of operation, the DVL runs on its own and can send a synchronization pulse to the SBG INS whenever a
new velocity measurement is available.
The measurement rate might be not constant depending on water conditions.
To operate in this mode, the DVL synchronization output must be connected to one of the available SBG INS SYNC
IN pins.

DVL Slave
In this mode of operation, the SBG INS generates a regular pulse signal to trigger each measurements of the DVL.
The measurement rate can be constant in this mode, driven by the SBG INS SYNC OUT rate.

Inertial Sensors Installation – 57

 Knowledge base – Knowledge Base

However care must be taken in setting up the SYNC OUT Rate to mach the minimum DVL's timing requirement,
based on DVL type, configuration, and maximum depth during the mission.
In other words, the DVL must be able to send a ultrasonic pulse, wait for the echo and process the data before the
next trigger shall occur.

Setting the right rate

SBG Systems recommends you to contact your DVL's manufacturer to know what allowable SYNC OUT
rates can be used depending on your system configuration and water environment.
As a rule of thumb, an update rate between 1Hz and 10Hz is feasible in shallow water conditions.

4.7 Lever arms & alignment calibration

4.7.1 Real-Time
A common problem in inertial sensor configuration and usage is setting-up properly the GNSS lever arm and other
external sensor installation parameters: A precise lever arm (within 1cm) is a key to ensure optimal precision. Angle
between inertial sensor an a dual antenna system should also be accurately measured within less than 0.05° to
ensure proper navigation performance.
However measuring angles and lever arms with such precision is generally not practically feasible, unless using a
total station. In challenging conditions, this expensive solution might not be able to account for small misalignment
between a short baseline dual antenna and a miniature inertial sensor (typical UAV setup).
To overcome these issues, SBG Systems have developed efficient algorithms and procedures to calibrate precisely
the installation lever arms and small angles. GNSS lever arms and DVL alignment are estimated through this
procedure.
This procedure can be performed in Qinertia post-processing software for optimal performance, or for high
performance navigation systems (Ekinox, Apogee, Navsight, Quanta), through the integrated web interface.
The procedure is slightly different in real time and in Qinertia, but the main steps are the following:
1. Enter a rough “a priori” configuration, with 10cm precision on each lever arm
2. Run a calibration sequence, with higher dynamics than usual operation to allow the EKF observing the
various lever arms and angles (moving typically through figure of eights).
3. Compute the refined calibration parameters using Qinertia or using the real-time calibration interface, and
apply these parameters in the sensor non-volatile memory
4. Use the sensor normally, with the precise lever arm setup to ensure high performance and robustness

Inertial Sensors Installation – 58

 Knowledge base – Knowledge Base

You can click on [calibration procedure to link when available] to see the see the real-time calibration procedure for
SBG Systems high performance units.

Ellipse series
On the Ellipse series, the default operation includes automatic lever arm estimation to allow easy setup.
Once refined parameters are obtained through Qinertia, it’s possible to stop this automatic lever arm
estimation to enable best performance.

Inertial Sensors Installation – 59

 Knowledge base – Knowledge Base

4.7.2 Post-Processing
As for real-time, it is possible to estimate lever arms and alignment using Qinertia post-processing software.
Qinertia offers the best solution available on the market to efficiently estimate mechanical installation parameters.
Please read the dedicated Lever Arm Estimation2 article in the Qinertia User Manual to get more details.
In Qinertia, this is easily done by telling Qinertia that the external sensor lever arms are not accurately measured. A
simple box in the GNSS configuration window must be unchecked:

The next time Qinertia processes the data, it will estimate the lever arms and possible alignment between the two
antennas. This estimation will appear in a window automatically once the processing is done and will show the
current settings compared to the estimated settings. You can then choose to use these new settings or not and to
launch a new processing. These steps can be done multiple times to converge towards the best lever arms which is
usually when the current lever arms are equal to the estimated lever arms.

2 https://support.sbg-systems.com/display/QD/Automatic+Estimation

Inertial Sensors Installation – 60

 Knowledge base – Knowledge Base

Of course as in real-time there are things to take into consideration when estimating lever arms and alignment in
post-processing. As in real time it is important to have dynamics to have a good estimation in 3D. It is also
important to do this while in PPK accuracy to measure these lever arms as accurately as possible.

4.8 Magnetic calibration

4.8.1 Overview and background

This article describes how to calibrate Ellipse sensors, when they are strapped on an object that affects the local
magnetic field.
Ellipse sensors are fully calibrated in factory. As for other sensors, magnetometers are fully temperature
compensated. However, magnetic sensors are very sensitive to their close environment (mainly the objects on
which they are strapped) : Some materials can generate magnetic fields that will be summed with Earth magnetic
field, and some other can distort the existing magnetic field. These effects will be measured by magnetometers and
will not be distinguished from the Earth magnetic field, and therefore error will occur in heading estimation. The
amount of this error can be really significant, reaching dozens of degrees in many cases.
SBG Systems has carefully chosen materials that do not disturb magnetic field in the Ellipse hardware. So the unit
itself should not disturb the magnetic field so much. Unfortunately, it is sometimes impossible to remove that kind
of materials in customer system. There are two kinds of distortions, which are described below.
Hard Iron distortions
This kind of distortion is caused by magnets, or anything that acts as a magnet. It is very easy to magnetize objects
like screws or nuts. Hard irons generate magnetic fields that are summed with the Earth magnetic field. Hard Iron
effect shifts the magnetic field measured by magnetometers by a constant offset, whatever the device orientation
is. Power supplies, which generate high current (several amps) and their associated wires, may also generate
magnetic fields.
Soft Iron distortions
This kind of distortions are caused by ferromagnetic objects which are placed in the vicinity of the device. Iron, and
some steels for example are ferromagnetic materials, but more generally, anything that sticks to a magnet is

Inertial Sensors Installation – 61

 Knowledge base – Knowledge Base

ferromagnetic. Ferromagnetic materials do not generate their own magnetic field. Instead of that, they react to
existing magnetic field (in our case Earth magnetic field). By this way, Soft Irons distort the local magnetic field in a
different way depending on the magnetic field direction. In addition to changing the magnetic field measurements,
it also tends to rotate it compared to inertial reference frame. That rotation can be quite significant, reaching tens
of degrees in some cases. This is why soft Irons are much more difficult to compensate for.

4.8.1.1 The most advanced calibration procedure available on the market

SBG Systems was the first company to introduce an easy to use hard & soft Iron calibration procedure in 2008 with
the IG-500 series. Since then, Hard and soft Iron calibration became a standard in the inertial sensor industry.
Always looking for best quality, SBG Systems stays a step ahead by using a new compensation of magnetic
coordinate frame alignment. When standard calibrations on market are subject to large and non-constant errors,
SBG Systems guarantee a reliable, yet easy to perform calibration.

4.8.1.2 What can be compensated for and what cannot

It is critical to distinguish disturbances that can be calibrated from disturbances than can not be.
Distortions that come from materials which are fixed with respect to the device coordinate frame can be
compensated for by the calibration procedure.
Distortions that are not fixed with the device, and move independently with respect to the device, or distortions
that change over time, cannot be compensated for and must be limited as much as possible.
Distortions of the magnetic field act generally very locally, so in practice keeping the device far enough from the
noise or time varying magnetic field might make the error negligible. Ideally, keep the device away from source of
disturbances from at least 3 meters.

Kalman filter
Thanks to its internal Kalman filter, the Ellipse is able to cope for short term external distortions.
Distortions that last up to a few minutes are handled without significant heading drift.

4.8.2 Procedure
The main goal of the procedure is to perform magnetic field measurements in many different orientations. This is
done by rotating the device in different orientations to measure the magnetic field. SBG Systems exclusive
algorithms are able to map the magnetic distortions in 3D and compensate for them. The more orientations will be
passed through, the best the results will be.
If you cannot rotate the device freely in all possible orientations, the calibration algorithm can easily cope with it,
and you will still get a good precision in orientations that were covered. However, a minimum of 9 significantly
different orientations is required to perform the calibration.
Two types of calibration are provided depending on the degree of freedom of the device:

Important Note
In order to get good results, it is important to keep away from at least 3 meters all external sources of
magnetic disturbances. Keep in mind that a building structure generally contains steel and other sources
of interference, as well as computers, chairs, desks, etc.

Inertial Sensors Installation – 62

 Knowledge base – Knowledge Base

4.8.2.1 3D Calibration
3D calibration procedure is the standard procedure and should be preferred as it will provide the best performance
in most applications. When user starts it, the device has to be rotated through the maximum amount of different
orientations. User has simply to rotate the device regularly at a relatively slow rate (< 250°/s rotations are
acceptable). Too fast movements may weaken results.
Ideally, the points should draw a complete ellipsoid shape. User has to try to cover a maximum of this ellipsoid
shape to get the best results. After calibration, the magnetic field norm should equal 1.0 in all orientations. In this
case, all points after calibration should be placed on a unit 3D sphere.

4.8.2.2 2D Calibration
Sometimes, it is not possible to move the device in 3D. A 2D calibration is intended for these cases. The procedure is
just to rotate the device through a horizontal circle. It is possible to cover less than a full circle, but best results are
achieved with a full 360° coverage. Before the calibration is performed, the magnetometers readings should form a
2D ellipse. After calibration, the magnetic field norm should equal 1.0 in all orientations. In the 2D case, all possible
magnetic fields should be placed on a unit circle.

The 2D calibration algorithm is specifically designed to make the best use of all available data, even with a
very limited motion. However, due to physical and mathematical constraints, it is not possible to fully
realign magnetic and inertial coordinate frames in that calibration mode. This can lead to a residual offset
on the heading measurements that depends on the magnitude of disturbances to correct.
There is no way to estimate this offset but you can read it by placing the device to a known heading.

The 2D calibration requires sensor position (Latitude, Longitude, Altitude) and date information in order to
work properly. It can be provided by “initial parameters” on an Ellipse-A product, or/and when the system
has entered in navigation mode for Ellipse E and N models.

4.8.3 Calibration methods

SBG Systems now supports two calibration methods that offer the same performance level. The choice of the
method to use depends on the user integration constraints.
• sbgCenter method which is the easiest way to proceed and should suit to most applications. It can use either
real time data, or previously recorded data. This requires a computer connected to the sensor, but it gives a
nice and powerful real-time feedback.
• On-board calibration method, that provides similar performance as the above one, but that do not require
any special piece of hardware to be executed. The Ellipse does it by itself when it receives corresponding
commands!

4.8.3.1 sbgCenter method

This method is advised if a computer can be connected to the device. It is very flexible and allows real time
calibration, as well as calibration with log files.

Inertial Sensors Installation – 63

 Knowledge base – Knowledge Base

Presentation of the calibration window

On the left of the window, a file list is available to include log files to reprocess or that has been previously recorded.
Multiple files can be used for a single calibration.
Then, two 3D plots are displayed:
• The first, “Before” shows all points in 3D as they are measured by magnetometers before any hard/soft iron
calibration. This display is updated in real time.
• The second one, “After” will show the same points, transformed by the calibration algorithm. This display
will only be updated once the Calibrate button is pressed.
These 3D plots can be easily rotated with the mouse, and you can also zoom in/out with the mouse wheel. When
you rotate a plot, the second one will also rotate so that the field of view remains the same in the two displays.
On the bottom, several buttons allow to:
• Reset calibration data
• Start/ Stop Real time acquisition
• Calibrate acquired data
• Save acquired data for future use
• Finish or cancel calibration
A few options will allow user to tune calibration behavior. These options can be used before or after the acquisition
has been performed, so it's possible to try new settings to see if the calibration is better performed in 2D or in 3D.
• Bandwidth, set to “High” by default, can be reduced to improve calibration robustness against noisy
magnetic field environments. In case of selecting “Normal” or “Low” bandwidth, it's then recommended to
move slowly the sensor during calibration (rotations < 100°/s). If observations are really noisy, then this
threshold will more likely reject the calibration when set to High than when set to Normal or Low.
• Mode can be set to 3D or 2D. 3D is the default one and should be used as long as it's possible to provide
sufficient motion. In fact, 2D calibration is a locally true correction which means that for large roll and pitch
angles and/or non negligible latitude changes, magnetic north would be rejected.
Finally a few boxes will inform user about the calibration procedure, its results, and some quality hints:

Inertial Sensors Installation – 64

 Knowledge base – Knowledge Base

• Points information: The number of collected points is displayed. The more points will be collected, the
better the calibration will be. The length of the current magnetic field (“After Norm”) is displayed once the
calibrate button is pressed. This length should always equal one in all orientation with a good calibration.
• Before deviation: This information is updated each time the calibrate button is pressed. It informs about the
distortion of the local magnetic field before calibration. The average deviation is representative of the
magnetic field. STD Dev defines the noise observed, and the maximum deviation informs about how noisy
the environment is.
• After deviation: This is the deviation of the magnetic field, when the full calibration algorithm is applied. The
average deviation is the most representative information. STD Dev defines the noise observed, and max
deviation represents the worse case of all measurements performed during calibration.
• Expected error: Finally, this box shows in term of degrees the expected magnetic heading error after the full
calibration. The STD error is the most representative information, and it comes with a color code (Dark green
optimal, Green is good, Orange is acceptable, Red is bad), which represents the confidence in the calibration
results. Max error represents the worse case of all measurements performed during calibration. If the max
error is high, it may come from a disturbed magnetic field during calibration.

Real Time Procedure

Before launching the calibration procedure, it is recommended to put the device in a place free of external
magnetic fields (more than 3 meters of any source of distortion).

Step 1: Start/Reset the procedure

Once you have started the Soft and Hard Iron Calibration Tool, the following window appears. When the device is
ready for calibration, you can click on the Start Acquisition button.

Inertial Sensors Installation – 65

 Knowledge base – Knowledge Base

Step 2: Rotate the device

Rotate the device ideally around each axis pointing downward and then upward for a 3D calibration. For a 2D
calibration, just make sure to rotate roughly in a plane (avoid too much roll/pitch motions). Rotate the device
slowly to get the best results. We recommend at least 8s to to complete a full 360° rotation around one axis.

Inertial Sensors Installation – 66

 Knowledge base – Knowledge Base

Step 3: Press the “Calibrate” button

If the calibration tool detects some issue during calibration, a pop-up window will appear to tell you what is wrong.
Possible errors might be having not enough motion observed during calibration, or alignment issues that come
from a too fast motion or too noisy magnetic field.

Step 4: Press OK to finish procedure.

The calibration data will be sent to the device and saved into internal Flash memory.

Inertial Sensors Installation – 67

 Knowledge base – Knowledge Base

4.8.3.2 On-board calibration

The on-board calibration method shares the same principle of operation as the sbgCenter one. So we suggest you
to first make use of the sbgCenter method to get familiar with the concepts involved in the magnetic calibration
before using the on-board method.
On-board and sbgCenter methods will provide similar performance. The on-board calibration has just a smaller
memory than a standard PC, so it's not able to store more than 1000 calibration points. Please note that this will not
affect performance as an intelligent system only stores relevant points in the internal memory.
The on-board calibration is using different low level commands to achieve the full procedure, and we will basically
run the same steps as sbgCenter method.
All the attributes mentioned above in the sbgCenter method and resulting from the calibration process will be
available with the On-board calibration method, except the “After Norm” value. Quality and confidence indicators
will be provided instead:

Quality indicator
• 0 Optimal: Magnetic field is very coherent with inertial motion after calibration
• 1 Good: Small magnetic field deviations from inertial motion have been detected. The magnetic calibration
should still provide accurate heading.
• 2 Poor: Large magnetic field deviations from inertial motion have been detected.
• 3 Invalid: No valid calibration has been computed. Magnetic calibration have failed or partially failed (Refer
to error pop-up).

Confidence indicator
• 0 High: Reported quality indicator can be trusted.
• 1 Medium: Few remarkable magnetic field points have been used.
• 2 Low: The data set used to compute the magnetic calibration was not meaningful.
Calibration should be considered applicable if Quality indicator is 0 OR 1, AND confidence indicator is 0.

Procedure

Initiate the calibration

The command SBG_ECOM_CMD_START_MAG_CALIB is sent to the Ellipse to start magnetic calibration. At this
point, user must define some calibration parameters:
• Bandwidth must be selected (high, medium, or low bandwidth). High is the default one; other options are
useful to cope with noisy magnetic fields.
• 2D or 3D calibration mode.

Compute calibration parameters

Using the command SBG_ECOM_CMD_COMPUTE_MAG_CALIB, the Ellipse will compute all calibration parameters,
and will return the calibration result, with advanced quality indicators, and indication about how to solve any
calibration issues.

Inertial Sensors Installation – 68

 Knowledge base – Knowledge Base

Store calibration parameters into the Ellipse internal memory

Once proper calibration parameters have been returned by the Ellipse, the user can decide to apply it. In order to do
so, the command SBG_ECOM_CMD_SET_MAG_CALIB must be sent with returned correction vector and matrix as
parameters.
Finally, the new configuration must be sent to flash memory for use at the next boot, using
SBG_ECOM_CMD_SETTINGS_ACTION command.

4.8.4 Advices and limitations

4.8.4.1 Cover a maximum of different orientations

Best results will always be given when a wide range of orientations is covered during calibration. In theory only a
few orientations would be necessary to map the whole magnetic field. But due to noise measurements, external
magnetic fields and other effects, it is always better to collect the maximum amount of data to get good results.

4.8.4.2 Limited degrees of freedom

If the device cannot be rotated freely in all orientations, the algorithm will still be accurate, but only in orientations
that were covered during calibration procedure.

Inertial Sensors Installation – 69

 Knowledge base – Knowledge Base

4.8.4.3 Keep away from any external sources of distortion

The magnetometers and the calibration algorithm cannot distinguish between distortions caused by external
magnetic fields and distortions caused by the object on which the device is strapped. Those external distortions can
be generated by the building structure, or Iron desks, computers etc.
It is therefore very important to move away the device from ferromagnetic objects within about 3 meters.

Inertial Sensors Installation – 70

 Knowledge base – Knowledge Base

This recommendation becomes even more important if the calibration is performed within a limited set of
orientations.

4.8.5 Specific applications

The Magnetic calibration for airborne applications(see page 71) page details how to perform the mandatory
calibration process on airborne applications.
The Magnetic calibration for marine applications(see page 74) page details how to perform the mandatory calibration
process on marine applications.

4.8.6 Magnetic calibration for airborne applications

When magnetometers are used as heading reference, a magnetic calibration is mandatory for normal sensor
operation. Different calibration methods are provided, depending on accuracy or ease of use requirement.

4.8.6.1 Light UAV calibration

As long as a UAV (fixed or rotary wings) is light enough to be held by a few people, a 3D calibration, made on the
ground is to be preferred. The basic procedure is the following:
1. Install the sensor as described in previous sections, and place the whole system away from external
magnetic disturbances (buildings, other vehicles, etc)
2. Press “Start acquisition” button on sbgCenter calibration window
3. Rotate the system as much as possible . The main point is to cover the whole flight profile, but a larger
amount of points, beyond the flight profile will provide even better results.
4. Press “ Calibrate ” and check calibration results. Press “ OK ” to finalize the calibration procedure.
5. Power cycle the sensor if you need immediate operation after calibration.

4.8.6.2 Airplanes, Helicopter and large UAV applications

In flight 3D calibration
This calibration will give the best results as it allows to map the magnetic field in real 3D so that magnetometers
readings are kept consistent even during turns and pitching.
In order to perform the calibration procedure, user can use the integrated sbgCenter calibration tool, or a data-
logger to store the “magnetic calibration data” outputted by the Ellipse during calibration procedure.

Procedure
Once the aircraft is in steady flight at a reasonable altitude, the goal is to cover different orientations which are
representative of the flight domain of the aircraft.
The calibration accuracy does not depend on any precise orientation (facing true North for example) and rather
depends on how many significantly different orientations have been covered. The calibration algorithms are able to
map the 3D magnetic field in orientation that have not really been covered during calibration; however, it is good to
cover the full flight domain to get the best results.

Inertial Sensors Installation – 71

 Knowledge base – Knowledge Base

For example an Extra 300 aerobatic airplane should get the best results by performing several representative
aerobatic maneuvers in different directions in order to get a good 3D coverage of the magnetic field. In the other
hand, a Cessna 172 private airplane could only perform high inclination eights to get optimal results.

Procedure tested on a private airplane

The following procedure has been tested with success on a piston private airplane.
The calibration starts in a steady flight. Two 360° turns will be performed decomposed in the following steps:

Step 1: Calibration Start. Press “start acquisition” button.

1. High bank right rolling – without turning.
2. High bank 120° left turn
Step 2:
1. High bank right rolling – without turning.
2. High bank 120° left turn
Step 3:
1. High Pitching: +20° then -20° then return to level flight
2. High bank right rolling – without turning.
3. High bank 120° left turn
Step 4:
1. High bank 120° right turn
Step 5:
1. High bank left rolling – without turning.
2. High bank 120° right turn
Step 6:
1. High Pitching: +20° then -20° then return to level flight
2. High bank left rolling – without turning.
3. High bank 120° right turn
Step 7: Calibration end. Press “ Calibrate ” button, then “ OK ” button write the calibration data to your sensor.

Inertial Sensors Installation – 72

 Knowledge base – Knowledge Base

Figure 1: Trajectory performed during calibration

Once these tests are done the calibration can end. It is not crucial to perform exact 120° turns, but the procedure
should perform rolling points at significantly different headings. In addition the pitching in the first turn should not
be performed at the same heading as the one done in the second turn.

This procedure can be easily transposed to rotor-craft. The procedure can be performed in a stationary
flight, by making several pitching and rolling maneuvers at different heading values. The goal is to expose
the sensor to as much orientations as possible.

Ground calibration (2D)

Although this method is not the most accurate, it's possible to calibrate the magnetometer on the ground, using the
“2D” calibration method.
The procedure is really simple and only requires a few steps on the ground to be performed:
1. Install the sensor as described in previous sections, and place the whole system away from external
magnetic disturbances (buildings, other vehicles, etc).
2. Place the aircraft on a horizontal platform. The aircraft must be kept horizontal (in its line of flight level). This
is the case with most tricycle landing gears airplanes, but this should be a concern with conventional landing
gears.
3. Calibration Start. Start the sbgCenter calibration tool and press “start acquisition” button.
4. Perform a 360° circle with the aircraft. The calibration mode has to be set on “2D”. The aircraft should be at
least 10m away from any metal building or other aircraft.
5. Calibration end. Press “Calibrate” and check calibration results. Press “OK” to finalize the calibration
procedure.

Calibration result examples

On the following screen-shots, it is possible to see that the calibration coverage is not a full 3D sphere but covers
significantly different orientations. The first screen shows an example of the calibration procedure explained above.
The second one shows a calibration only performed with a simple “8” performed.

Inertial Sensors Installation – 73

 Knowledge base – Knowledge Base

Please click the link for more information about Magnetic calibration(see page 61).

4.8.7 Magnetic calibration in marine applications

When magnetometers are used as heading reference, a magnetic calibration is mandatory for normal sensor
operation. Different calibration methods are provided, depending on accuracy or ease of use requirement.

4.8.7.1 Light Ship calibration (ASV, ROV, AUV)

As long as a vessel is light enough to be held by a few persons (especially unmanned vehicles), a 3D calibration,
made on the ground is to be preferred. The basic procedure remains the same, and you should just rotate the
system in as much orientations as possible.

4.8.7.2 Large ship calibration

In case the vessel is a heavy ship and it is not possible to move it by hand, it will be necessary to calibrate the
magnetometers while cruising. The goal will be to collect data in every direction, so you will have to make a 360°
with the ship (the turn radius and speed does not matter).
You should prefer a 3D calibration if your boat can heel, if not then a 2D calibration should be done.
The following procedure should be followed for good performance:
1. Install the sensor as described in previous sections, and place the whole system away from external
magnetic disturbances (buildings, other vessels , etc)
2. Press “Start acquisition” button on sbgCenter calibration window
3. Navigate in an 8 shape pattern, so you will be able to capture points in the Y-axis while heeling slightly on
both sides. It doesn’t need to be a critical angle, 20° would be enough for instance, it has to be
representative of the usual behavior of the ship.
4. Check that the 3D method is used (in case of very large ship, or when the roll/pitch angles could not be
changed significantly during calibration, a 2D method can be used). Press “ Calibrate ” and check
calibration results. Press “ OK ” to finalize the calibration procedure.
5. Power cycle the sensor if you need immediate operation after calibration.

Inertial Sensors Installation – 74

 Knowledge base – Knowledge Base

Please click the link for more information about Magnetic calibration3

4.9 Specific features regarding the application

The installation of the system (IMU, GNSS antenna, etc...) has to be done in accordance with your application. You
can find here, recommendations to use our products in specific conditions.

• Airborne(see page 75)

• Land(see page 78)
• Marine(see page 82)

4.9.1 Airborne

4.9.1.1 Mechanical installation

Airborne applications assume a 3D motion. The INS sensor can be located anywhere in the aircraft, considering the
following recommendations:
• Sensor is rigidly fixed to the aircraft frame : avoid locations that may be bending like wings
• Sensor is not moving in regard to other equipment (antennas, LIDAR, etc...)
• Sensor is far from strong vibrations sources : avoid placing the sensor directly on the engine mounts.
• Sensor should be protected from high temperature variations.
• When relevant (if magnetometers are used), place the sensor away from magnetic disturbances like strobe
lights, high current equipment, radios or moving parts like landing gears.

Note
SBG Systems IMU are designed to handle vibrations without specific care. Nevertheless in case of highly
vibrating environment, a mechanical vibration isolation might be required for proper operation. Silicon or
wire dampers can be used for that purpose.

Aircraft reference frame

The vehicle coordinate frame is defined as follows :
• X axis points to the front of the aircraft
• Y axis points rightward.
• Z axis points downward.

Note
The sensor can be placed in any orientation in the aircraft. When IMU axes do not match exactly with the
aircraft coordinate frame, the rough and fine alignment parameters should be corrected through the
configuration interface to realign the IMU and aircraft coordinate frames.

3 https://support.sbg-systems.com/sc/kb/latest/inertial-sensors-installation/magnetic-calibration

Inertial Sensors Installation – 75

 Knowledge base – Knowledge Base

4.9.1.2 GNSS setup considerations

When installing your INS with a GNSS aiding, you will need to install the GNSS antennas with a clear view of sky, and
fixed with respect to the IMU.
The GNSS lever arms shall also be measured, which are the signed distance, expressed in the aircraft coordinate
frame, FROM the sensor center of measurements, TO the GNSS antenna.
We usually require these measurements to be precisely performed, within 1cm accuracy.

Note
It's generally not practical to measure with such precision the lever arms, so SBG Systems developed lever
arm calibration that enables you to measure rough lever arm estimation (10cm precision) and let the tool
refine those measurements.

The GNSS lever arms should be lower than 10m to minimize induced errors.

Single antenna installation

Single antenna installation is possible when regular dynamics are experienced and allow to converge heading
angle. Typical heading performance will degrade during straight lines and will re-converge during dynamic phases.
A single antenna installation with GNSS lever arm is shown below:

Inertial Sensors Installation – 76

 Knowledge base – Knowledge Base

Dual antenna installation

Dual antenna may be required if dynamics are expected to be low during extended periods of time (eg. navigation
with more than 5min of straight lines). When using the INS in a dual antenna setup, the heading will remain stable
and precise in all conditions. Heading can also be initialized in static conditions.
Dual antenna systems installation will require special care in order to obtain optimal performance :
• The antennas must be fixed with respect to the to the inertial unit
• Same antenna type should be used
• Same cables with identical lengths must be used for both antennas. If Splitter are used make sure that
they are adapted and with the same characteristics
• Both antennas must have the same view of sky. Typically avoid placing antennas on each side of the rudder
or parts that could mask a significant part of the sky.
• If the antenna model does not have integrated ground plane, a 10 cm diameter ground plane must be
added for both antennas.
• If antennas are not permanently installed on the aircraft, antennas reference marks (usually the connector
position) should be mounted in a repeatable fashion in order to guarantee antennas phase center stability
from mount to mount and minimize changes to heading misalignment angle.
Both GNSS antennas lever arms should be measured accordingly.

Inertial Sensors Installation – 77

 Knowledge base – Knowledge Base

4.9.2 Land

4.9.2.1 Mechanical installation

Land applications assume a 2D motion with (depending on the motion profile) limited to no lateral velocity. The INS
sensor can be located anywhere in the vehicle, considering the following recommendations:
• Sensor is rigidly fixed to the vehicle frame
• Sensor is not moving in regard to other equipment (antennas, LIDAR, etc...)
• Sensor is far from vibrations sources.

Note
SBG Systems IMU are designed to handle vibrations without specific care. Nevertheless in case of highly
vibrating environment, a mechanical vibration isolation might be required for proper operation. Silicon or
wire dampers can be used for that purpose.

Vehicle Reference Frame

The vehicle coordinate frame is defined as follows:

Inertial Sensors Installation – 78

 Knowledge base – Knowledge Base

• X axis points to the front of the car

• Y axis points rightward
• Z axis points downward.

Note
The sensor can be placed in any orientation in the vehicle. When IMU axes do not match exactly with the
vehicle coordinate frame, the rough and fine alignment parameters should be corrected through the
configuration interface to realign the IMU and vehicle coordinate frames.

Primary lever arm

Once the sensor is installed in the vehicle, the center of Rotation of the vehicle should be identified. This is most
often located along the rear axle, at the ground level.
The Primary Lever Arm which is the signed distance in the vehicle frame, FROM the IMU, TO the Center of Rotation
should be measured within a 5 cm accuracy.

Inertial Sensors Installation – 79

 Knowledge base – Knowledge Base

4.9.2.2 GNSS setup considerations

When installing your INS with a GNSS aiding, you will need to install the GNSS antennas with a clear view of sky
(usually on the roof of the vehicle), and fixed with respect to the IMU.
The GNSS lever arms shall also be measured, which are the signed distance, expressed in the vehicle coordinate
frame, FROM the sensor center of measurements, TO the GNSS antenna.
We usually require these measurements to be precisely performed, within 1 cm accuracy.

Note
It's generally not practical to measure with such precision the lever arms, so SBG Systems developed lever
arm calibration that enables you to measure rough lever arm estimation (10cm precision) and let the tool
refine those measurements.

Single antenna installation

Single antenna installation is possible on all automotive applications with the only limitation that the heading will
not be observed while stationary.
A single antenna installation with GNSS lever arm is shown below:

Inertial Sensors Installation – 80

 Knowledge base – Knowledge Base

Dual GNSS antenna placement

Dual antenna may be required if dynamics are expected to be low during extended periods of time. When using the
INS in a dual antenna setup, the heading will remain stable and precise in all conditions. Heading can also be
initialized in static conditions.
Dual antenna systems installation will require special care in order to obtain optimal performance:
• The antennas must be fixed with respect to the to the inertial unit
• Same antenna type should be used
• Same cables with identical lengths must be used for both antennas. If splitters are used make sure that
they are adapted and with the same characteristics
• If antennas are not permanently installed on the vehicle roof, antennas reference marks (usually the
connector position) should be mounted in a repeatable fashion in order to guarantee antennas phase center
stability from mount to mount and minimize changes to heading misalignment angle.
• Both antennas must have the same view of sky: typically place the antennas on the roof of the vehicle.
• Baseline of at least 1 meter between both antennas is recommended for best performance
• If the antenna model does not have integrated ground plane and is not placed on a metal roof, a 10 cm
diameter ground plane must be added for both antennas.
Both GNSS antennas lever arms should be measured accordingly.

Inertial Sensors Installation – 81

 Knowledge base – Knowledge Base

4.9.3 Marine

4.9.3.1 Mechanical installation

The INS sensor can be located anywhere in the vessel, considering the following recommendations:
• Sensor is rigidly fixed to the frame
• Sensor is not moving with regard to other equipment (antennas, sonar, LIDAR, etc...)
• Sensor is far from vibrations sources
• Sensor is not exposed to salty water, unless sub-sea casing (IP-68 is not corrosion-proof)
• When relevant (if magnetometers are used), place the sensor away from magnetic disturbances like high
current equipment, radios or moving parts.

Note
SBG Systems IMU are designed to handle vibrations without specific care. Nevertheless in case of highly
vibrating environment, a mechanical vibration isolation might be required for proper operation. Silicon or
wire dampers can be used for that purpose.

Vessel Reference Frame

The vessel coordinate frame and positive rotations for Euler angles are defined as follows :
• X axis points to the front of the vessel (bow)
• Y axis points to the right (starboard)
• Z axis points the bottom (keel).

Note

Inertial Sensors Installation – 82

 Knowledge base – Knowledge Base

The sensor can be placed in any orientation in the vessel. When IMU axes do not match exactly with the
vessel coordinate frame, the rough and fine alignment parameters should be corrected through the
configuration interface to realign the IMU and vessel coordinate frames.

4.9.3.2 GNSS setup considerations

When installing your INS with a GNSS aiding, you will need to install the GNSS antennas with a clear view of sky, and
fixed with respect to the IMU.
The GNSS lever arms shall also be measured, which are the signed distance, expressed in the vessel coordinate
frame, FROM the IMU center of measurements, TO the GNSS antenna.
We usually require these measurements to be precisely performed, within 1cm accuracy.

Note
It's generally not practical to measure with such precision the lever arms, so SBG Systems developed lever
arm calibration that enables you to measure rough lever arm estimation (10cm precision) and let the tool
refine those measurements.

The GNSS lever arms should be lower than 10m to minimize induced errors.

Dual GNSS Antenna Placement

With a dual antenna setup, the INS will be able to keep a stable and precise heading as long as it has a clear GNSS
signal. Heading can also be initialized in static conditions.
Dual antenna systems installation will require special care in order to obtain optimal performance :

Inertial Sensors Installation – 83

 Knowledge base – Knowledge Base

• The antennas must be fixed with respect to the IMU

• Same antenna type
• Same cables with identical lengths must be used for both antennas. If splitters are used make sure that
they are adapted and with the same characteristics
• If antennas are not permanently installed onboard, antennas reference marks (usually the connector
position) should be mounted in a repeatable fashion in order to guarantee antennas phase center stability
from mount to mount and minimize changes to heading misalignment angle.
• Both antennas must have the same view of sky. Typically avoid placing antennas on each side of a structure
or parts that could mask a significant part of the sky
• Baseline of at least 2 meters between both antennas is recommended for best performance
• If the antenna model does not have integrated ground plane, a 10 cm diameter ground plane must be
added for both antennas.
Both GNSS antennas lever arms should be measured accordingly.

Single antenna installation

Due to the very particular motion of a boat, a system with a single GNSS antenna is recommended only with an INS
that supports magnetometers (SBG Ellipse series).
A single antenna installation with GNSS lever arm is shown below:

Inertial Sensors Installation – 84

Knowledge base – Knowledge Base

Inertial Sensors Installation – 85

 Knowledge base – Knowledge Base

5 Inertial Sensors Operation

This section provides details about SBG Systems inertial sensor behavior during operation.

• SBG EKF modes of Operation(see page 86)

• Real time operation with cm level accuracy(see page 91)
• Ultimate accuracy with post-processing(see page 94)
• Time, synchronization and event management(see page 97)

5.1 SBG EKF modes of Operation

5.1.1 sbgEkf Introduction

Thanks to a modern processing architecture, SBG Systems inertial systems run a real time Extended Kalman Filter
(EKF). The coupling between GPS/GNSS and inertial sensors allows GNSS data correct any inertial drift, while
keeping high frequency navigation outputs, with excellent performance.
More than just a direct EKF implementation, the implemented algorithms include advanced error models and
wrong measurement detection to ensure that best navigation performance is provided at any time.
A modular design allows a wide range of aiding sensors to be connected to the INS. GNSS, Odometer, DVL and other
aiding sensors can be connected to further enhance navigation performance.
In addition, the Extended Kalman filter is able to estimate some user entered parameters to further improve
accuracy, such as GPS lever arm, odometer's gain, and others.
Specialized motion profiles and error models provide optimal options and tuning for each application, and each
aiding equipment.

5.1.2 Basic principle

Inertial sensors (accelerometers and gyroscopes) provide very accurate short term motion measurements but suffer
from drift when integration time becomes long. Some other systems such as GNSS receivers or odometer provide
low frequency measurements that can be fooled by jamming, or short term measurements errors, but these sensors
provide good performance over long term.
The basic idea behind the Kalman filter is to take the best of each sensor, without drawbacks. A high frequency
prediction (also called propagation) step uses inertial sensors to precisely measure motion and navigation data.
When aiding data (GPS position, Odometer data or DVL reading for example) becomes available, the Kalman filter
will use it to correct the current state and prevent drift.
As aiding measurements are made at a lower frequency than the prediction step, a small jump can be observed
after a correction is applied. This jump should be really small in normal operating conditions.
A covariance matrix maintains up to date each estimated parameter error. When there is no measurement
available, estimation error tends to increase; when a new measurement is received, this error will decrease. This
covariance matrix is also used to handle the “link” between each estimated parameters.
Besides the EKF, a sensor manager is implemented to check aiding measurements and reject bad ones.
To summarize the EKF operation, the following diagram shows how IMU and external sensors are used

Inertial Sensors Operation – 86

 Knowledge base – Knowledge Base

5.1.3 Mode of Operation

The Kalman filter will run into several computation modes depending on situations.

5.1.3.1 Uninitialized mode

This mode is observed at startup only. Prior to the first attitude computation, the filter is identifying the average
direction of gravity using accelerometer. This mode assumes low accelerations so best initialization is achieved
when the device is powered up stationary or at constant speed. If the INS is powered up during motion, the full
accuracy may be reached within a few minutes after startup.

5.1.3.2 Vertical Gyro mode

Once roll and pitch angles are initialized, the EKF filter is running in a vertical gyro mode, where only roll and pitch
angles are estimated. This mode uses a vertical reference and internal gyroscopes to estimate orientation.
Therefore, heading angle is freely drifting. Ship motion data is provided but may have degraded accuracy in
dynamic environments.
In this mode the performance is dependent on the dynamics of the vehicle. Performance can be affected during
high acceleration maneuvers.

5.1.3.3 Heading Rough alignment procedures

This section only applies in case the unit is configured to operate in AHRS or INS modes of operation.
While being operated in “vertical gyro” mode, the EKF is continuously trying to make a first heading angle
alignment using different procedures. These procedures have some constraints and the following table explains
how they are used and in which situations:

Inertial Sensors Operation – 87

 Knowledge base – Knowledge Base

Method Availability Constraints - Remarks

Magnetic When Magnetic This mode is common for entry level applications, with no high
Heading heading aiding is precision requirement. If magnetometers are enabled, this signal is
enabled. available at startup.
A good magnetic field must be available for proper operation.
Other methods are recommended for high precision applications.

GNSS Dual When GNSS dual In case the setup allows two GNSS antennas to be installed, this
antenna antenna Heading is method will be the most convenient. It uses GNSS true heading
Heading activated as an aiding provided by a dual antenna GNSS receiver to align heading.
input.
Good GNSS condition are preferred for initialization in order to avoid
multi-path errors at startup. A minimum accuracy of 1° for the GNSS
true heading has to be reached, to be used in the solution.

Kinematic In Automotive and This method allows heading alignment for applications where heading
alignment Airplane and motion is mostly aligned with travel direction (course over ground). It uses
profiles. GNSS velocity, considering that preferred direction of travel is forward.
The device must drive/fly in forward direction, at least at 3.0 m/s.

Free In Helicopter, UAV, This method uses relative velocity to define a heading. This allow any
Kinematic marine and pedestrian motion, in any direction unlike the traditional kinematic alignment
Alignment motion profiles methods.
In order to operate properly, the device should experience linear
acceleration or accelerated turns during a time frame of 5 to 10
seconds. The alignment precision will be driven by the amount of
dynamics experienced.

5.1.3.4 AHRS mode

Once the heading rough alignment has been performed, the full orientation can be estimated by the EKF. The
vertical reference is still stabilizing the roll and pitch angles. Heading is also stabilized thanks to the dual antenna or
magnetic aiding. Position and velocity are freely drifting and cannot be considered as valid in this mode.
As for the Vertical gyro mode, the performance can be affected by high dynamic maneuvers.

5.1.3.5 Position and velocity initialization

As for heading alignment, the system continuously tries to initialize the position and velocity using GNSS inputs
when operated in AHRS mode.

5.1.3.6 Full Navigation mode

In this mode, the EKF provides the full navigation outputs: Orientation, absolute position and velocity are
estimated.

Inertial Sensors Operation – 88

 Knowledge base – Knowledge Base

The Full navigation mode can use all available sensors input to maintain the best solution, even in case of short
GNSS outages.

AHRS - Navigation Mode

AHRS units behavior slightly differ from one product line to another: The Ellipse-A cannot accept aiding
data so it only runs in the AHRS mode. However, higher performance motion reference units (MRUs)
accept an external GNSS input to allow navigation mode to be used internally. This improves orientation
output performance. Only orientation outputs remain available in this case.

5.1.4 Heading Observability

This section is important to understand how a GNSS aided INS is able to track an accurate heading.

5.1.4.1 Using Single Antenna GNSS

In the most simple setup, with only an INS aided by a single GNSS antenna, the heading is not always observed.
In particular, when de device is subject to static conditions or at constant speed, only roll and pitch angles are
accurately corrected, and heading can show some drift (reported on the standard deviation outputs). As soon as the
device is subject to acceleration, the EKF will also stabilize heading angle.
In case the reported Yaw angle standard deviation becomes high or the error increases, then doing some dynamic
maneuver will improve accuracy. This will help in airborne application with UAVs for example where only a single
antenna is available.

5.1.4.2 Using Dual Antenna GNSS

The heading observation is greatly improved in low dynamic conditions when coupling the INS with a dual antenna
GNSS system. In such condition, the INS will be able to provide accurate heading in all conditions, and will cope
with GNSS heading outages in case of multipath environment. Dual antenna heading also provides a very accurate
heading angle, often required by survey applications.
This option is particularly effective in marine applications which have low dynamics and a high precision
requirement.

5.1.4.3 Automotive applications

In automotive applications, it's often possible to assume that there is no lateral velocity. This assumption allows the
filter to optimize its performance. Thus heading becomes fully accurate as soon as the vehicle is driving.
Using a dual antenna heading can further improve heading performance in case the vehicle is regularly in static
condition.

5.1.4.4 Using Magnetic Heading

In many applications such as airborne or sometimes in marine environment, the magnetometer can provide an
efficient heading aiding input as long as it is calibrated and used away from magnetic interference. The Magnetic
Heading is converted to True heading by using the position and date and the World Magnetic Model 2020.
Magnetic heading is a cost effective way to observe heading in low dynamic conditions but can be subject to
degradations in difficult magnetic conditions. Its accuracy is limited to 1° approximately.

Inertial Sensors Operation – 89

 Knowledge base – Knowledge Base

5.1.5 Improving Navigation Performance

5.1.5.1 Lateral velocity constraints

When the "Automotive" motion profile is selected, the EKF assumes that lateral velocity is 0. In addition to improve
the heading performance, this will greatly reduce the position drift in challenging conditions such as urban
canyons.
Regular turns are required to constrain the velocity in all directions.

5.1.5.2 Odometer Aiding

In addition to the GNSS aiding, all INS models provide an odometer input which can greatly improve performance in
challenging environments such as urban canyons for automotive applications. The odometer provides a reliable
velocity information even during GNSS outages. This increases significantly the dead reckoning accuracy.
Compared to using only the velocity constraints, the odometer aiding provides useful velocity information that will
be very effective in case of dead reckoning in straight lines or slight turns.
Our products support:
● Quadrature output or compatible odometers with forward and reverse directions.
● CAN vehicle velocity messages (fully configurable) for setup with direct interface with vehicle’s ODBII connector

Odometer integration is made really simple as the EKF will finely adjust the odometer's gain and will
correct residual errors in the odometer alignment and lever arm.

5.1.5.3 Doppler Velocity Log Aiding

In many marine or underwater applications, the DVL is a good choice to improve navigation when GPS is not
available. DVL has been fully coupled with the EKF to provide full navigation performance in both bottom tracking
and water layer conditions. No calibration is required as the EKF will automatically adjust alignments and gain
parameters.
The fusion of DVL data with the EKF can provide very accurate and reliable underwater position data in real
conditions. A carefully chosen mission pattern such as a lawn mower one can also dramatically limit the position
error growth.
In addition to the Kalman filter integration with DVL, the inertial sensor can store and output back the DVL
messages (PD0) for water profiling applications.

5.1.5.4 ZUPT Mode

The EKF is able to automatically use “Zero Velocity Updates” (ZUPT) in some motion profiles. When the sensor stops
moving, the Kalman filter detects the zero velocity condition and uses that information to correct the states, and
then limit the position drift.

Inertial Sensors Operation – 90

 Knowledge base – Knowledge Base

5.2 Real time operation with cm level accuracy

This article covers integration aspects required for reaching centimeter level accuracy in real time applications. For
general GNSS technology knowledge click on this link: GNSS(see page 29).

Post-processed applications
Note that some application don't require the navigation solution to be computed in real time. In such
case, it might be relevant to investigate how post-processing could simplify your setup. The following
article (see page 94)covers the high precision applications in post-processing.

5.2.1 RTK
RTK (Real-Time Kinematic) positioning is a technique used in GNSS navigation to have a more accurate position up
to 1cm in real time (1cm + 1ppm). RTK uses two receivers: one base and one rover that are located in the same area
(up to 20km typically).
The major advantage of RTK is that is is able to cancel all the common errors involved in the GNSS algorithms. In
particular, the atmospheric errors caused by ionosphere and troposphere are cancelled when rover and base are
close enough. As the distance between rover and base increase, the residual atmospheric errors becomes more
important and will impact the RTK solution (typically an error increase of 1ppm of the distance between base and
rover).
This means that with a 20km baseline, the RTK accuracy will be 1cm + 1ppm x 20km = 3cm.

Inertial Sensors Operation – 91

 Knowledge base – Knowledge Base

Relative position vs Absolute position

Note that RTK technology provides a relative positioning to the base station to the cm level. In order to
obtain a centimeter level absolute accuracy, the base station used must be precisely located to the
centimeter level as well.

5.2.1.1 Typical RTK setup

In order to enable RTK operation, your system will be composed of three major components:
• The Rover receiver, located on the vehicle. This is typically the GNSS receiver that is integrated into your
navigation system. This receiver must allow RTK computations (permissions).
• The Base Station (or Reference Station) is a static GNSS receiver on the ground that sends corrections to
the Rover via a wireless link. The base station can be installed by the surveyor, it can be part of an already
existing network of base stations, or it can even be virtually created thanks to a network of base stations.
• A way to communicate between the Rover and Base Station - typically a UHF, or 4G GSM modem.

5.2.1.2 RTK corrections formats

There are standard different formats that are used for RTK corrections:
• RTCM2.x => Older format from the ages of DGPS, now replaced by RTCM3.x
• RTCM3.x => Most interoperable correction format, supported by all our products
• CMR => Older Trimble proprietary format
• CMR+ => Multi constellation Trimble correction proprietary format

5.2.1.3 Installing a base station

To install your own RTK base station you will need the following:
• A GNSS receiver and antenna that will be used as the base station
• A radio system (one for the base and one for the rover) for wireless connection to the Rover
The base station is going to send the RTK corrections wirelessly via the radio to the rover.
If the absolute position accuracy is an important aspect, the base station position must be precisely surveyed using
a PPP technique.

5.2.2 Network RTK (NRTK)

A base station network can be used instead of installing your own base station. Some networks are free to use,
some require a fee. They can be accessed online via an NTRIP client.

Inertial Sensors Operation – 92

 Knowledge base – Knowledge Base

5.2.2.1 NTRIP
An NTRIP client is used to connect to a server via the Internet. Once connected, it will receive corrections from a
base station nearby, or from a virtual base station created using multiple base stations in the surrounding.

5.2.2.2 VBS / VRS

A Virtual Base Station or Virtual Reference Station (VRS) is a base station that doesn’t exist physically, but is
created virtually using several physical base stations. This is used usually when the closest base stations are too far
to be used accurately. A VBS usually requires the position of the Rover in order to be created as close as possible to
it for the best accuracy.

Inertial Sensors Operation – 93

 Knowledge base – Knowledge Base

5.2.3 PPP
Precise Point Positioning can be used in case no base stations can be used nearby. This is typically used in offshore
applications where there is no stable land to install a static base station.
The drawback with using PPP is that the solution takes time to converge (around 20 minutes) so it should not be
used in harsh environments.
PPP services include Omnistar, Marinestar, and others.

5.3 Ultimate accuracy with post-processing

5.3.1 Tight Coupling PPK

PPK (Post processed Kinematic) and is a combination of 2 things:
• The addition of a base station in post processing even if one was not used in real time and post processing
applied to the GNSS data.
• A forward/backward and merge done on the final inertial solution.

Inertial Sensors Operation – 94

 Knowledge base – Knowledge Base

These 2 things can improve significantly the data logged in real-time, especially for applications where there are
GNSS outages like for example in bridge surveying.
For more information on real-time vs post-processing you can follow Real time vs Post-Processed Operation(see page
31).

5.3.1.1 What is needed

In order to have a PPK solution, we need a post-processing software such as Qinertia, raw IMU data from the INS,
raw data from the GNSS receiver and of course all the data should be accurately timestamped.

5.3.1.2 PPK in Qinertia

The PPK or Post processed Kinematic uses the same algorithm as RTK but in a post-processed environment. This
mode of operation has the major advantage of improving the performance, maximizing the centimeter level
availability in challenging conditions. This is possible because there is no risk of RTCM corrections outage in post-
processed applications, but also to the nature of the forward/backward/merge processing.
The other key advantage in PPK is that the real time setup is simplified as there is no need to feed the RTCM
corrections to the rover GNSS. These corrections just need to be logged separately at base station level.
The PPK can also be performed using one of the 7000+ open access base stations available around the world,
further simplifying the real time setup.

Inertial Sensors Operation – 95

 Knowledge base – Knowledge Base

Single base station in Qinertia

In Qinertia, you can either load your own base station data if you had a base station installed during the survey, or
you can choose one from a network of available base stations.

Whatever solution you choose, it is possible to do a PPP computation (Precise Point positioning) on the base station
to calculate its position accurately. This allows you to make sure the position you entered manually is accurate or to
calculate the base station position if not entered correctly in real time for ultimate accuracy.

VBS in Qinertia
The Virtual Base Station concept, developed initially for network RTK providers, can be extended to post-
processing, using the freely available base stations.

Inertial Sensors Operation – 96

 Knowledge base – Knowledge Base

In case the mission to post-process is too wide for a single baseline RTK processing, or when base stations are too
far from the trajectory, it is then possible to use a Virtual Base Station network to enable centimeter level accuracy
of the rover position.

Qinertia VBS implementation is particularly efficient in processing VBS networks, with a distance from the rover to
the bases easily reaching up to 100km or more. This feature greatly extends the availability of PPK.

5.3.2 Post-processed PPP

Precise Point Positioning is an alternative solution in case no base stations were available nearby, and it can change
some offshore data to a 10cm accurate position.

5.4 Time, synchronization and event management

When working in a system with different devices that continuously communicate/log time-stamped data, the
synchronization of the devices' clocks is crucial to be able to consolidate the data.
This article addresses the clock management within the SBG system. To begin, we will provide an in-depth
understanding of the different calculation delays within an SBG INS. Following that, we will see how an SBG INS
synchronizes its own clock with UTC time. Finally, we will explore the different tools to synchronize SBG products
with other devices.

Inertial Sensors Operation – 97

 Knowledge base – Knowledge Base

5.4.1 INS output latency

The output latency is an important aspect in real time control applications, where a higher output latency could
degrade control loops performance. SBG INS embedded software has been designed to minimize output latency:
once sensor data are sampled, the Extended Kalman Filter (EKF) performs small and constant-time computations
before the outputs are generated.
Typically the observed output delay is less than one millisecond.
Note that CAN Logs might be received after all serial outputs because the CAN protocol cannot guarantee the
output delay in case of traffic on the CAN bus.

Good to know
The processing latency should be added to the data transmission latency if you want to get total delay.
This transmission latency vary from one interface to another.
For instance, a 50 bytes message sent on a UART interface at 115200 bps will take 4ms for complete
transmission. Consider higher baudrates to minimize output latency.

5.4.2 Events

5.4.2.1 Event inputs

Inertial sensors can include up to 5 synchronization inputs that can be used for different purposes:
• Event triggered logs: All pulses received generate events that can generate specific Logs output. Any output
log can be triggered by an event pulse.
• Event Marker: An event marker log can be sent each time a pulse is received in order to time mark each
event.
• PPS input: When connected to an external GNSS system, the PPS signal is used to realign and synchronize
internal clock to GPS clock.
• Other aiding input timestamping: If a specific aiding sensor generates pulses that time stamp the following
output, the corresponding event input can be used for data synchronization.

Event triggered logs

In this mode of operation, the inertial device runs on its own clock to sample sensor data and compute the inputs.

Inertial Sensors Operation – 98

 Knowledge base – Knowledge Base

Once a new output is ready to send, the system will simply check if an event input was received during last
processing loop, and will send the output messages accordingly.
This mode of operation may induce an output jitter of up to 5ms due to the synchronized internal and host clocks.

Warning
SBG Systems sensors handle up to 200Hz input for output log triggering. In case of higher frequency
events, only the last received event will be considered for message triggers.

Event markers handling

Event markers allows you to precisely measure the time of events such as camera shutter time, up to the
microsecond precision. Once received, events inputs are precisely timestamped and stacked in the system and a
marker output message will be sent during the next message output window.
This log will include all events received during previous loop.

Warning
SBG Systems sensors handle up to 1kHz event Marker’s input. Sending more than 1KHz events may
overload the internal CPU.

5.4.2.2 Event output

The INS can be used to synchronize other devices by generating pulses on a synchronization output pin.
These pulses can be generated based on the following modes:
• Main loop divider: an event is sent at the beginning of an INS loop (at the sensor sample time). A divider can
be configured to reduce its frequency. For example, If the divider is set to 4, the pulse output frequency will
be 200Hz / 4 = 50Hz.
• PPS: an event is sent at the beginning of an INS loop (at the sensor sample time), at 1Hz frequency. If the
system is time synchronized with a UTC input, this output is provided at each top of a second in UTC time.
• Virtual odometer: an event is generated each time a configurable distance has been traveled (feature
available on our High Performance products such as Ekinox, Apogee, Quanta and Navsight).

5.4.3 Synchronizing with external devices

There are three main solutions to synchronize with external equipment.
• PPS + time messages output. This mode is very similar to common GNSS receivers operation and
enables less than 1µs accuracy. See the Event output section of this document.
• Network Time Protocol (NTP) allows for synchronization with a 1 millisecond accuracy on local network.

Inertial Sensors Operation – 99

 Knowledge base – Knowledge Base

• Precision Time Protocol (PTP) allows for synchronization with an accuracy of 150 nanosecond to 1µs . It
however requires specific Hardware so is not always usable.

5.4.3.1 Network Time Protocol (NTP)

NTP is a distributed service that synchronizes the clocks of equipment to a coordinated universal time UTC based
on TCP/IP networks.
The protocol uses a hierarchical (master-slave topology) of time source, with each level called a Stratum. The
highest level (Stratum 0) are the atomic clocks, such as the ones in GNSS satellites. Additional stratums can be
added to synchronize equipment on a specific network.
The NTP is only implemented in Stratum 1 to distribute time to other devices. Time is marked as not synchronized
until an internal UTC time is available.
NTP uses the UDP protocol, on port number 123.

If you need further details on the NTP protocol you can check the NTP wikipedia page4.

5.4.3.2 Precision Time Protocol (PTP)

Like NTP, PTP is a protocol used to synchronize clocks throughout a TCP/IP network. However it is built to offer a
substantially better precision than NTP.
To reach this level of precision, the protocol has a Hardware dependency, and is therefore not applicable on all
devices. All SBG inertial devices featuring an Ethernet connection enable the PTP time server, with a measured
average time accuracy of 150ns.
The PTP implementation generates a clock marked as faulty in case UTC time is not available internally.
Once UTC time is available, our PTP system can operate as Grand master clock (or passive mode if a higher
precision time reference is available on the network).
PTP operates uses the UDP network protocol, on port number 319 and 320.
SBG PTP implementation provides timestamps in TAI time scale.

If you need further details on the PTP protocol you can check the PTP wikipedia page5.

4 https://wikipedia.org/wiki/Network_Time_Protocol
5 https://wikipedia.org/wiki/Precision_Time_Protocol

Inertial Sensors Operation – 100

 Knowledge base – Knowledge Base

6 Technology insights
This section covers specific topics in a detailed way.

• INS benefits over MRU in Hydrography(see page 101)

• How to compare IMU(see page 103)
• Antenna Characteristics: How to choose the correct one(see page 113)
• UART Baud Rate and Output Rate(see page 115)

6.1 INS benefits over MRU in Hydrography

In Marine Survey applications, it is critical to compensate the SONAR measurements from the ship motion. In the
same time, a position input is required to geo-locate the point cloud. For many years, the traditional way to do so
was to integrate two separate components : one Motion Reference Unit (MRU), and one GNSS receiver, capable of
RTK processing to achieve centimeter precision. If this solution is suitable for many basic hydrographic
applications, it may struggle in more challenging conditions like bridge survey or operation in rough sea.

6.1.1 Principle
An MRU (Motion Reference Unit) also called AHRS (Attitude Heading and Reference Systems), is a dynamic
inclinometer. It uses gyroscopes to detect the rotations and accelerometers to detect gravity, and compute its
orientation to provide Roll and Pitch. It may use the Earth Magnetic Field or Earth Rotation rate to detect its heading
(in such case we will talk about a Gyro-compass). It also uses its accelerometers to detect the waves motion and
provide the Ship Motion (Heave Surge and Sway) measurement. MRU performance is optimal in low dynamic
applications, and can be significantly affected by dynamics.
An INS is a Full Inertial Navigation System that will use GNSS aiding as primary aiding source to complement the
inertial measurements. Thanks to the inertial and GNSS coupling, the INS is able to output position even in case of
GNSS disturbances. Moreover the INS orientation output is insensitive to dynamics, making it perfect for operation
in challenging conditions.

6.1.1.1 MRU Outputs

Technology insights – 101

 Knowledge base – Knowledge Base

6.1.1.2 INS Outputs

Advantages
A quick comparison between MRU with an independent GNSS receiver and INS clearly indicates that if the
navigation data are necessary, then INS will be a preferable solution rather than MRU with GNSS.

MRU with GNSS INS

Affordability + ++

Ease of configuration + ++

Performance ++ +++

Outputs + +++

6.1.2 INS Measurement's improvement

In an INS, the Heave algorithm can be aided by the GNSS velocity to keep optimal accuracy even in rough sea,
where a single MRU would purely depends on its accelerometers. This means that even if we are looking purely into
attitude and ship motion, the GNSS will already bring a performance gain.
Now if we also consider the position is necessary, then using an INS (rather than MRU+GNSS) will greatly improve
the position performance. The inertial will indeed compensate for the GNSS inaccuracies and outages.
A typical example is a bridge survey application. When going under a bridge, an independent GNSS receiver position
would be inaccurate or even lost. Using an INS instead of MRU + GNSS will allow a continuous position instead. The
next 2 pictures show the comparison between a GNSS only (in yellow) compared to a INS position (in green), while
cruising under a bridge during a survey.

Technology insights – 102

 Knowledge base – Knowledge Base

6.2 How to compare IMU

Sensors on the market will have datashteet that will include all or part of the following specifications. It matters to
know the meaning of each of these specifications if present on a datasheet as it is a decision maker to select an IMU.

6.2.1 Full Scale

It represents the measurement range were the sensor has been calibrated (for instance +/- 2g on accelerometer or
+/-500°/s on a gyroscope), it can usually go a bit further than the maximum specified values, but the accuracy of the
measurement is not guaranteed. For instance the next graph shows the gyroscopes saturates at 500°/s, which
means any faster rotation will not be measured by the sensor.
It is not recommended to select a measurement range smaller that the expected motion.

Technology insights – 103

 Knowledge base – Knowledge Base

6.2.2 Scale Factor Stability

Scale factor error are measurement inaccuracies proportional to the applied motion. An advanced sensor thermal
calibration can minimize this kind of errors. SBG Systems specification of scale factor error includes thermal and
long term stability errors.
The next graph shows an ideal sensor (in blue) would report the actual input of the motion, while a realistic sensor
(in red) would report a slight offset on the output, compared to the actual motion on the input.

Technology insights – 104

 Knowledge base – Knowledge Base

6.2.3 Non-linearity
The non-linearity is similar to the scale factor: the output of the sensor is not corresponding to the actual input it
should measure. However instead of having an output with proportional error depending on the full scale (in red) it
will actually varies in a more complex way (in green).
Non linearity errors can be minimized by advanced calibration procedures.

Technology insights – 105

 Knowledge base – Knowledge Base

6.2.4 One Year Bias Stability

A one year bias specification is a really important information in the sensor datasheet. It reflects how much the
sensor will drift over its life time. In order to obtain such specification, SBG Systems runs a sensor accelerated
ageing process including thermal cycles, power cycles, shocks and vibrations stress.
This parameter reflects what part of the sensor bias cannot be calibrated, and how much the navigation filters will
need to compensate. At each start the INS will go through a warm-up process to estimate the sensor biases. The
lower this value, the easier it will be for the INS to estimate and compensate the bias, lowering its warm-up time.

Technology insights – 106

 Knowledge base – Knowledge Base

6.2.5 Velocity/Angular Random Walk

The Velocity Random Walk for accelerometers, or Angular Random Walk for gyroscopes represents the noise of
the sensor. While the bias corresponds to the long-term noise, the Random Walk corresponds to the short-term
noise. This noise, while integrated, will cause the Random variations of the output.
The Velocity Random Walk is usually expressed in m/s/√hr or m/s/√s or g/√Hz
The Angular Random Walk is usually expressed in °/√hr or °/√s
This is especially important for pointing applications that requires an accurate Roll and Pitch.

Technology insights – 107

 Knowledge base – Knowledge Base

6.2.6 In Run Bias Instability

The bias of an IMU changes over time while powered on, due to various effects such as temperature, and
mechanical stress This is why the INS constantly needs to re-estimate this bias with external aiding, such as GNSS,
odometer, DVL, etc., combined with vehicle dynamics.
In run Bias instability is a kind of magic number for all MEMS IMUs, with the lower value always being
targeted. However, this value doesn't give much insights about the actual sensor performance. Especially because:
• This value is typically measured at constant temperature
• The Bias Instability is never specified with associated time slot (tau). If the bias instability is obtained at very
low time slots (for instance tau < 200s), then this information is meaningless because the bias will quickly
diverge.

Technology insights – 108

 Knowledge base – Knowledge Base

6.2.7 Allan variance

The Allan Variance is a complex diagram, showing for various time scales how good is the sensor.
Allan variance is typically measured at ambient and control temperature so user should still be careful when
comparing Allan variances.
When the high frequency noise is "Gaussian", which is typically the case, one can measure the Angular/Velocity
random walk at a time scale of 1second.
The Bias Instability is the bottom of the curve.

Technology insights – 109

 Knowledge base – Knowledge Base

What is more important is how the sensor behave after reaching its bias instability. While the best MEMS sensors
will tend to stabilize, lower grade sensors or uncalibrated sensors will quickly ramp up, showing a sensor bias drift.

6.2.8 Vibration Rectification Error

It is generally provided on highly qualified systems, while most IMUs just do not specify what happens in vibrations.
The Vibration Rectification Error gives the expected bias for a given level of vibrations on accelerometers or
gyroscopes.
For instance if the VRE is 1 °/h/g² this means we can expect a maximum of 1°/h of bias for 1g of random vibrations.

Technology insights – 110

 Knowledge base – Knowledge Base

6.2.9 Sampling Rate

The Sampling Rate is the frequency at which the measurements are taken on the sensor's input. The sampling rate
divided by 2 is the Nyquist frequency, which gives the maximum observable frequency of a signal we are measuring.
This parameter matters if we wish to observe higher frequency motions, for instance to achieve vibration
monitoring or fast motion compensation.

6.2.10 Bandwidth
The Bandwidth is maximum frequency you can take a reliable acceleration or angular velocity reading. This will
directly represent which frequency of motion you will be able to measure, as any signal with a frequency higher
than the bandwidth will be attenuated.
This point should be considered for any application that involves vibration or fast motion analysis.

Technology insights – 111

 Knowledge base – Knowledge Base

6.2.11 Resolution
The Resolution is the smallest increment measurable on a sensor.
This parameter is of less importance, as usually much lower than the noise of the sensor. The sensor angular or
velocity random walk will be the actual decisive factor as this parameter gives the actual system noise level during 1
second averages.

Technology insights – 112

 Knowledge base – Knowledge Base

6.2.12 Orthogonality
The Orthogonality represents the physical misalignment between the sensor of different axis.

6.2.13 Cross-Axis Sensitivity

This is the sensitivity of a sensor to a vibrations applied to a perpendicular axle. Ideally it should be minimal as each
axle should be independent from each other.

Gyro-G
Gyroscopes can have a bias when under acceleration, this error is called gyro-g effect. This is typically estimated
and compensated during calibrations.

6.3 Antenna Characteristics: How to choose the correct one

In this article we will talk about the different specifications that are important when choosing an antenna and what
makes an antenna compatible with your GNSS receiver, or better than another one.
We will use the Ellipse-N/D internal receiver as an example here. You can find below the specifications for it:

Technology insights – 113

 Knowledge base – Knowledge Base

6.3.1 Connection:

6.3.1.1 Connector type:

There are different connector types available depending on the antennas and sometimes they are not the same as
the connector type on the receiver side. This is easily fixed with an adapter. For example, the Ellipse-N/D uses SMA
female connectors because they are smaller, but the Tallysman TW3972 uses TNC female connectors so most of the
time a TNC cable male male will be used with it. An SMA male to TNC female adapter will allow you to connect the
TNC cable to the antenna on one side and to the Ellipse on the other side.

6.3.1.2 Input Impedance:

This one is easy as 50 Ohm is an industry standard for coaxial cables and power amplifiers, so most coaxial cables,
GNSS receivers and antennas will use 50 Ohm.

6.3.2 Power supply:

6.3.2.1 Active vs Passive:

Active antennas are passive antennas that also have an integrated Low-Noise Amplifier (LNA). Because of this,
active antennas need a power supply that is usually fed through the coaxial RF cable by the GNSS receiver. SBG
Systems embedded GNSS receivers require the use of active antennas.

6.3.2.2 Voltage:
The power supply that an antenna will accept is usually within a certain range. For example, the Tallysman TW3972
antenna will accept 2.5 to 16 VDC so it will be compatible with the Ellipse receiver that outputs 3V here, but also the
Ekinox or Apogee receivers that output 5V.

Technology insights – 114

 Knowledge base – Knowledge Base

6.3.2.3 Current draw:

Some antennas will draw more or less current compared to another antenna, and this current draw can depend
also on the voltage used. The Tallysman TW3972 requires 24mA. So it is important to make sure the receiver can
provide this much current. The Ellipse-N or D can output up to 30mA so it is compatible.

6.3.3 Signal:

6.3.3.1 Amplifier gain :

The antenna LNA gain is the ratio of input to output power.
The gain of the antenna should be within a range that is compatible with the GNSS receiver. For example the
Tallysman TW3972 will have a gain of 37dB, and the Ellipse will be compatible as long as the gain is between 17 and
50 dB, so this antenna is compatible with the Ellipse.

6.3.3.2 Constellations :
A constellation is a set of satellites that work in the same network. GPS is the American constellation, Glonass is the
Russian one, etc. A constellation is usually usable anywhere on earth, the GPS constellation is American but is
available worldwide to everyone for example.
It is important to use an antenna that can use constellations compatible with the receiver. If the antenna is only
compatible with GPS, then even if the receiver is compatible with GPS, Glonass, Beidou and Galileo, it will only use
the GPS constellation which is limiting. And vice versa if the antenna is multi-constellation but the GNSS receiver is
only compatible with GPS for example. The best choice would be an antenna that is compatible with the same
constellations as the GNSS receiver.

6.3.3.3 Frequency band:

A same constellation will use different frequencies, also referred to as bands (L1, L2, L5 or L6). When using a dual
antenna GNSS receiver for heading computation it is important to use L1/L2 antennas at the least because L2 is
required for this. With an L1 only antenna the receiver won’t be able to compute a heading.

6.4 UART Baud Rate and Output Rate

A UART (Universal Asynchronous Receiver/Transmitter) interface is a widely used serial communication protocol for
transferring data between devices, such as microcontrollers, computers, and various peripherals. UART
communication is asynchronous, meaning it doesn't rely on a shared clock signal between the sender and receiver.
Instead, it uses a predefined baud rate to determine the speed at which data is transmitted.
In this article we'll define what the baud rate and output rate are exactly, and help you understand how to select
the correct baud rate depending on the quantity of data being output to avoid saturating the serial output.

Technology insights – 115

 Knowledge base – Knowledge Base

6.4.1 Definition

6.4.1.1 Bits and Bytes

Bits are the smallest increment of data and hold a single binary value of 0 or 1.
A byte is a group of 8 bits.

6.4.1.2 UART frame

A UART frame typically includes 1 start bit and 1 or 2 stop bits (usually 1) during transmission, and sometimes even
a parity bit that is used for error checking and can be set to odd, even, mark, space, or no parity.
So for each byte transmitted over UART, 10, 11 or 12 bits are transmitted per frame depending on the combination
of stop bits and parity bit.
A transmission configuration of 8N1 for example will mean that for each byte transmitted there is 1 start bit (implicit
here), 8 data bits, no parity bit, and 1 stop bit, for a total of 10 bits.

6.4.1.3 Baud
Baud is the number of symbols transferred per second. 1 baud is equivalent to one bit per second.
The following formula can be used to calculate how many bits per second are transferred.
Baud = number of bytes x total bits per frame x output rate of message (in Hz), where total bits per frame = data bits,
+ start bit + stop bit + parity bit if used.

6.4.1.4 Baud rate

The Baud rate, expressed in bits per second (bps) is the rate of symbols transferred across a transmission medium
per second.
The standard baud rates are the following: 4800, 9600, 19200, 38400, 57600, 112500, 230400, 460800, 921600.

It is important to have the same baud rate configured on each side of the communication.

6.4.1.5 Output rate

The output rate is the number of times a message is output per second. For example outputting a specific message
at 50Hz means that we will output the same message 50 times per second, so every 0.02 seconds.

Technology insights – 116

 Knowledge base – Knowledge Base

6.4.2 Selecting the correct baud rate for a specific amount of data output per
second
To select an appropriate baud rate, it is important to know how many bits will be output per second. Let's take the
example of our sbgECom protocol. Each message is encapsulated in a 9 byte frame described in our firmware
manual6. If we want to output the EKF Euler message that is 32 bytes long (the description can also be found in the
firmware manual), then we'll have the following:

Field SYN SYN MSG CLAS LEN DATA CRC ETX

C1 C2 S

Size (bytes) 1 1 1 1 2 32 (EKF Euler here) 2 1

Description Syn Syn Messa Messa Lenth of Payload data 16 bit End
c. c. ge ID ge DATA CRC of
wor wor Class section fram
d d e

Value 0xF 0x5 - - - - - 0x33

F A

Total size 41
(bytes)

So to output this message we'll need to be able to output 41 bytes per second.
Let's assume we are using the common UART configuration of 8N1 as defined above. In this case, the total bits per
Frame is 10 (8 data bits + 1 start bit + 1 stop bit).
If we want to output this 41 bytes message at an output rate of 50Hz at 8N1, then we'll have: 41 bytes x 10 bits per
frame x 50 Hz = 20500 bits per second or 20500 bauds.
So in order to transmit 20500 bits per second, we will need a baud rate of at least 38400. If a lower baud rate like
19200 was selected, then the port would saturate.

Recommended baud rate

In general, a common baud rate to use is 115200, but it might not be enough for the amount of data
needed for some applications. For post-processing where we require IMU data at 200Hz, you will need to
use a baud rate if at least 230400.

When outputting multiple messages, you can do a sum of the bits transmitted per second over UART, for each
message at their specific output rate, to determine how many bits total are transmitted per second.

6 https://support.sbg-systems.com/download/attachments/17993338/Ellipse%20Ekinox%20Apogee%20Series%20-%20Firmware%20Manual.pdf?
api=v2&modificationDate=1620229949007&version=1

Technology insights – 117

 Knowledge base – Knowledge Base

6.4.3 Calculating the amount of bytes per second that can be output for a specific
baud rate
Instead of calculating which minimum baud rate would be needed for a specific amount of data being output, we
can calculate the total amount of bytes per second that can be output at a specific baudrate before saturation.
Let's assume we are using the common UART configuration of 8N1 here too and have to use a baudrate of 115200.
We can use the following formula:
Bytes per second = baud rate / total bits per frame.
So, at a baud rate of 115200 bps with an 8N1 configuration, we can send 11520 bytes per second.

6.4.4 Length of cable vs baud rate

The baud rate selected will have an impact on the maximum length of cable we can use in the application. The
higher the baud rate, the shorter the cable will need to be for the signal to arrive properly on the other end. You can
use the table below to determine what length of cable to use depending on the baud rate selected with RS232 or
RS422.
RS232/RS422 : A standard defining the signals between two devices (the signal names, their purpose, voltage levels,
connectors and pinouts).

Length of Cable (m)

Baud rate RS232 RS422

28 800 bps and below 100 250

57 600 bps 50 250

115 200 bps 25 250

230 400 bps 12.5 250

460 800 bps 5 100

921 600 bps 2 50

6.4.5 How to check for port saturation

The sbgCenter can be used with SBG Systems products to check for port saturation. The transmit status in the
Interfaces window will show a "ok" status when everything is good, and a "no" status when the port is saturating.
This status is taken directly out of the system if the SBG_ECOM_LOG_SYSTEM_STATUS message is output.
For example in the image below we can see the Port A transmit is ok but the Port E transmit is saturating.

Technology insights – 118

 Knowledge base – Knowledge Base

It is also possible to check if any product is close to saturating its output with our sbgDataLogger. You can see below
the input status will be green if everything is ok and will change to orange or red depending on the amount of bytes
received per second compared to the baud rate selected.

Technology insights – 119

Knowledge base – Knowledge Base

Technology insights – 120

 Knowledge base – Knowledge Base

7 sitemap

sitemap – 121