Ph.D.

Research Proposal
Doctoral Program in Energy and Environment

Collision avoidance in space environment

PhD Candidate
Karthick Dharmarajan
karthickdharman@yahoo.com
Matricola:1824175
SAPIENZA UNIVERSITY OF ROME
 Abstract

Space environment is a new field in environmental science which is gaining relevance.

Spacecraft orbital collisions due to another spacecraft and orbital debris has been getting
increasingly common in space environment. With the number of spacecraft in space set to
get even more higher there is an urgent requirement for a new spacecraft collision
detection and avoidance system. Currently there are more than 18000 artificial satellites in
space and more than 35000 objects larger than 10cm with another million objects large
enough to cause serious damage to the satellites. The first step is in tracking all the objects
in space that has the potential to damage a spacecraft. When it comes to a possible collision
between two satellites it is important to consider that such a collision would not be certain
but is defined in terms of probability associated with the intrinsic uncertainty in predicting
satellite orbits. In fact, both the initial sensors’ measurements and propagation (in
mathematical terms- numerical integration of the nonlinear Newton equation including
perturbation effects) are affected by errors. These errors propagate in time causing large
deviations in orbit paths and in process occasionally bringing them in path of another
satellite. Hence availability of an effective collision avoidance system becomes mandatory.
The collisions in space not only affects the colliding objects but also create lot of debris that
will add to the debris already present in space and hence possess a greater probability for
even more collisions. This debris cloud makes spacecraft operation in the entire altitude
range dangerous. And if the satellite that is getting damaged is an operating satellite then it
results in even more losses. If the operating satellite can be controlled and if there is enough
information available to perform a collision avoidance manoeuvre, the orbit of the satellite
can be altered slightly to minimize the likelihood that a collision could happen amounts.
With an early warning, the chance of collision can be lessened by simply adjusting the time
when a station keeping, or an orbit maintenance operation is performed. The technique
discussed here is to develop collision avoidance manoeuvres for satellites exploiting the
effect of Solar radiation pressure and drag perturbations without the use of excessive of
thrust for both Low Earth Orbits and High Earth Orbits (LEO and HEO), and also taking into
account the required constraint of limited consumption common to all spacecraft operation.
Table of Contents

INTRODUCTION 4

STATE OF THE ART 5

COLLISION DETECTION 5
COLLISION AVOIDANCE 5

RESEARCH OBJECTIVES AND APPROACH 6

WORK PLAN AND IMPLICATIONS 7

CONCLUSIONS 7

REFERENCES
 INTRODUCTION

Space environment is a branch of science that seeks to understand and address conditions
existing in space that affect the operation of spacecraft. The space environment is a gravity
less dangerous place where the satellites sent from Earth operate and perform their
functions whether they are civilian, scientific or military. The effects of spacecrafts can arise
from radiation, Space Debris, Meteorite impact, atmospheric drag and from ionized particles
present in space. Collision avoidance is a set of procedures performed to make sure that an
operating or non-operating spacecraft does not collide with another natural or artificial
satellite. The process of collision avoidance starts even before the spacecraft is launched. A
detailed analysis is performed for any possibility of collisions with another satellite and
debris. Then once the satellite is in orbit there is always a risk of satellite collision because
there is always an uncertainty in data provided. This uncertainty is propagated for huge
number of orbits and the collision probability is monitored. Once the probability of collision
exceeds a certain threshold a collision avoidance manoeuvre must be decided and
performed. Aside from the obvious task of reducing the risk of collision the manoeuvre must
also be executed in a way that reduces the propellant use. The worst case scenario for
collisions in space is the Kessler syndrome (also called the collisional cascading, or ablation
cascade), proposed by the NASA scientist Donald J. Kessler in 1978,in which once the
amount of debris in space reaches a critical mass a chain reaction is started that gives rise to
more debris which causes additional collisions which in turn causes creation of more debris.
This problem is of even greater concern in the future with plans like STARLINK from SPACE-X
to send 4425 satellites in a constellation to space and recent efforts from companies like
Amazon and Boeing. These projects mean the space is going to be more crowded than ever
before. At the very end, space debris issue could become a show-stopper for any new
launch, quickly and unfortunately limiting or ending the space exploitation.
In the following paragraphs I will present the state-of-the-art techniques about tracking of
the debris and satellites, which is providing the mandatory information about numbers of
objects and uncertainty in their kinematic state measurements by currently available
techniques. I will also provide some basic information about collision avoidance technique.
The main part of the essay is devoted to the concepts of the proposed research, attempting
to illustrate the main lines which could and should be followed to improve the collision
avoidance capabilities. Then I will briefly define a possible work plan I would follow, if
admitted to the PhD course, to tackle the problem in the three-year period.
Conclusions and references will end my essay.
 STATE OF THE ART
COLLISION DETECTION:
New guidelines have been proposed for satellite with low earth orbits to have a lifetime of
not more than 25 years. These satellites must perform a deorbiting manoeuvre and return
to lower altitudes where they burn during re-entry. Still there is a possible risk of satellite
collision.The TLE (Two-Line Elements) catalogue maintained and provided by the
USSTRATCOM (US Strategic Command) is the only publicly available and reasonably
accurate orbit information. The information given by the USSTRATCOM is of dual use. It can
used by both the military and the public. Despite the fact the information does not provide
any uncertainty data and the quality is insufficient for accurate monitoring of spacecraft
information the TLE is mandatory for any spacecraft monitoring. Many space agencies are
strongly involved with the issue. The data from the TLE’S is used to propagate the orbit
using the SGP4 MODEL which approximates the effects of perturbations like drag, solar
pressure, gravitational perturbation etc. Even the most theoretically accurate propagation
of orbit models is not enough for proper predicting of collision probability and collision
avoidance manoeuvre execution. Various techniques have been proposed to improve the
observed data. GSOC(German Space operations centre) proposed techniques including
radar tracking campaign and also studied the effects of solar flux on orbit uncertainty .The
numerical orbit propagator is used to produce osculating ephemeris data. The TLE is
determined from a best fit to the generated osculating trajectory. In the paper New
approaches has been proposed using real time propagation ” A REAL-TIME ORBIT
SATELLITES UNCERTAINTY PROPAGATION AND VISUALIZATION SYSTEM USING GRAPHICS
COMPUTING UNIT AND MULTI-THREADING PROCESSING” proposed a new approach based
on Monte Carlo with orbiting objects based on a Monte Carlo method with multi-threading
(CPU) and multi-stream(GPU) processing for the application of multiple satellite orbits
estimation. Various techniques using unscented Kalman filter, Sparse grid quadrature are
also proposed.

COLLISION AVOIDANCE:
The science of collision avoidance is a recent one. Most medium to large sized satellites
have a certain ΔV allotted to station keeping requirements typically in the range of 20m/s
per year. Recently it has been proposed to have a ΔV in the range of 2m/s for collision
avoidance manoeuvres. So, it is extremely important to perform the collision manoeuvres as
infrequent and also as fuel efficient as possible. A number of factors have been taken into
account to achieve this including the time of executing the said manoeuvre. Efforts should
be taken to keep the collision probability below 10e-4. The latest techniques include the
development of an Analytic Model predictive controller (MPC). Reactive collision avoidance
control, performed without preplanning, is used to avoid collisions while reconfiguring. But
reactive avoidance control is not as fuel-efficient as preplanning control as they only kick in
once the collision is identified. MPC anticipates system behaviour to proactively correct for
observed disturbances thus reducing collision probability before any chances of collision
even takes place. One of the techniques that is proposed currently is to use attitude control
for non-circular satellites to slowly change the trajectory of satellites. Currently most anti-
collision avoidance systems depend on providing a good collision probability data and no
particular collision avoidance system has been commonly used.

RESEARCH OBJECTIVES AND APPROACH

Modern collision detection algorithms while good are still not good enough for orbit
propagations over long periods of time. The size of the uncertainty ellipsoid increases
rapidly making a planned collision avoidance manoeuvre even more difficult. A new
machine learning approach using the propagated orbit data from TLE’S and SGP4 model
could be introduced and by comparing the numerical and analytical approaches to orbit
propagation. A cost function with penalty could be introduced for the results obtained. And
by reducing the penalty with each iteration a more accurate orbit could be propagated.
Collision avoidance manoeuvres using perturbations such as Solar pressure, Drag and
Gravitational are studied. The perturbations cause a slight change in the semi axis of the
satellite. Drag and Gravitational perturbations are prominent in low earth orbits and solar
pressure is prominent in higher orbits. These disturbances in turn are enough to shift the
satellite slightly enough to avoid collisions. Usually since the collision avoidance system is
performed using propulsion manoeuvres the above method is more fuel efficient. Typical
formation flying approaches take into account these perturbations and have control to be
maintained in formation. A model predictive control that uses the drag and solar
perturbations is developed for formation flying spacecraft that actively regulates the control
in case of an advancing collision and varies the control or switches off the control that is
required to be in formation so that collision is avoided. The accurate data obtained from the
orbit propagation will help us to identify an optimal control manoeuvre that can be
implemented several orbits early before the expected collision takes place. This accurate
data is very important because control using natural perturbations is a very slow process
and we need to have the knowledge of impending satellite collision several orbits before the
actual collision takes place. Thus, the main objectives of the research are to develop a more
accurate satellite propagation algorithm and use the effect of natural propagations to create
a model controller utilising natural perturbations to perform collision avoidance
manoeuvres for satellites in formation. Another challenge is getting back the satellites in
formation once the collision avoidance manoeuvre has been performed. A study whether it
is possible to use the perturbations to get the satellite back in its initial orbit shall be
performed.
 WORK PLAN AND IMPLICATIONS

The progress of study involves first taking a list of satellites and debris large enough to be
identified in space around earth orbit. Any object greater than 1cm is considered potentially
dangerous and considered for collision avoidance manoeuvre. The TLE data obtained is
obtained and propagated with SGP4 models. A numerical and analytical analysis is
performed. Various other propagation models are also considered for the study and
investigated. Common types of formations in satellite formation flying is discussed and their
future importance is studied. The increasingly important role of satellite formations means
that a collision strategy for satellites in formation is important. We introduce here a strategy
based on control using orbital perturbations. A detailed study of all perturbations is
performed for the different orbits mentioned in the study and the effect of perturbations on
each individual satellite and formations is done and then the data is then applied in the
control algorithms. The results of the study will help in us obtaining a more detailed and
accurate model of orbit propagation for spacecrafts in space and gives us solutions for the
problem of collisions for satellites in formation.

CONCLUSIONS

With the number of spacecrafts in space rapidly increasing the threat of Kessler effect is
increasing every day. With it there is a growing importance of studies about the collision
avoidance systems. The proposal presented is to help achieve a more accurate monitoring
and propagation of satellites and with it potentially new collision avoidance technique for
individual satellites as well as satellites in formation that could save lot of fuel but also be
more accurate and precise. Space as an environment need to be protected and the first step
is to make sure to avoid creating more man-made debris around the Earth.
 REFERENCES
1. Collision Avoidance Operations for LEO Satellites Controlled by GSOC (SpaceOps
2010 Conference) by S. Aida, M. Kirschner, M. Wermuth and R. Kiehling(German
Space Operations Center) ,82234 Wessling, Germany
2. Space traffic management: towards safe and unsegregated space transport
operations (Progress in Aerospace Sciences) by Samuel Hiltona, Roberto Sabatini,
Alessandro Gardi, Hideaki Ogawa, Paolo Teofilatto
3. Analytic Model Predictive Controller for Collision-Free Relative Motion
Reconfiguration(JOURNAL OF GUIDANCE, CONTROL, AND DYNAMICS Vol. 35, No. 4,
July–August 2012) by Luke Sauter and P. Palmer (University of Surrey, Surrey,
England GU2 7XH, United Kingdom)
4. Collision Avoidance for Satellites in Formation Flight(JOURNAL OF GUIDANCE,
CONTROL, AND DYNAMICS Vol. 29, No. 5, September–October 2006) by G. L. Slater,
S. M. Byram and T. W. Williams (-University of Cincinnati, Cincinnati, Ohio 45221-
0070)
5. Energy Suboptimal Collision-Free Reconfiguration for Spacecraft Formation Flying
(JOURNAL OF GUIDANCE, CONTROL, AND DYNAMICS Vol. 29, No. 1, January–
February 2006) by Cornel Sultan ,Sanjeev Seereram and Raman K. Mehra
6. General Method for Calculating Satellite Collision Probability(JOURNAL OF
GUIDANCE, CONTROL, AND DYNAMICS Vol. 24, No. 4, July–August 2001) by Russell P.
Patera (The Aerospace Corporation, El Segundo, California 90245-4691)
7. Satellite Formation-Containment Flying Control with Collision Avoidance(JOURNAL
OFAEROSPACE INFORMATION SYSTEMS) by (Liangming Chen, Yanning Guo, and
Chuanjiang Li and Jing Huang