VOCODERS

The coder is a hardware circuit (chip) or software program that converts the
spoken word into digital code and vice versa.
Two types of speech coding techniques are
(a) PCM/ADPCM (16-64 kbps) and (b) Parametric coding (vocoders)…

A vocoder in general is a speech

analyser and synthesizer.
Vocoders may be classified on
the basis of bit rate as follows:
Medium rate : 8–16 kbps
Low rate : between 8 kbps and
2.4 kbps
Very low rate : below 2.4 kbps

-- Human speech producing organs

 Few Speech Parameters
The pressure impulses are commonly called pitch impulses and the
frequency of the pressure signal is the pitch frequency or
fundamental frequency .

Typical pitch impulse sequence

Variations in pitch frequency

with time
• The pitch impulses stimulate the air in the mouth cavity and for
certain sounds also the nasal cavity.
• When the cavities resonate, they radiate a sound wave, which is the
speech signal. Both cavities act as resonators with characteristic
resonance frequencies, called formant frequencies .
Speech production

(a) Human speech

production
modelling (b)
Equivalent
synthetic speech
production blocks
Quality versus Bit rate for Speech Coders

Vocoders may be used in musical instruments, television and films, robots, or talking
computers.
 Types of Vocoders (As per approach)
The two basic speech coding methods for data rates between 4.8
kbps and 16 kbps are analysis and synthesis (AaS) and analysis by
synthesis (AbS).

• AaS---In the AaS approach, an analyser in the transmitter analyses

the original speech and extracts a set of parameters that represent
some kind of source filter model. These parameters are then
transmitted to the receiver, where a synthesizer reconstructs the
speech based on the received parameters.

• AbS---In the AbS approach, an analyser and a local synthesizer are

introduced in the transmitter. The synthesized speech is now
available in the transmitter for analysis.
 Different Vocoders
• Channel vocoder
• Formant vocoder
• Cepstrum vocoder (forms the basis for subband
coding)
• Voice-excited vocoder
• Linear predictive coding vocoders (forms the basis for
hybrid coders)
 Basic Analysis-Synthesis Function in a Channel
Vocoder
Channel vocoders are parametric, frequency domain vocoders,
among the analysis and synthesis systems
• channel vocoders encode and multiplex the samped envelope of the
speech sg for a num of frequency bands.
• They use bank of filters or dsp to divide signal into several sub bands.
• Output of env follower is a ctrl sg or vltg proportional to strength of
that slice
• Therefore o/p of analyser is a set of slowly varying ctrl vtgs that
constitute the code or analysis of spectrum of the sg.
• Set of Second audio sg called replacement, carrier or excitation sg fed
to bpf at synthesizer.
Formant Vocoders (Parametric)
Transmits formant information

Voiced speech
(a) Sample (b) PSD

Unvoiced speech
(a) Sample (b) PSD
• Voiced & unvoiced sgs
• Voice sgs air pressure from the lungs forces the normally closed vocal cords to open & vibrate. Vibrational
frequencies vary from 50-400 hz and form resonance at odd hormonics. These resonance peaks are called
formants.
• Unvoiced no vibration. They appear more noiselike periodicity.
• Formant vocoders transmit formant info i.e, position of the peaks of spectral envelope instead of sending
samples of entire power spectrum signal.
• Since speech sg info contained mainly in formants, vocoder that can predict position and BW of formant can
achieve very high quality at a low rate.
• Operate at 1kbps
• Typically formant vocoder must be able to indentify at least 3 formants for representing a speech sound.
• Drawbacks:
1) lower info rate
2) greater distortion
Cepstrum Vocoder
• Its also parametric vocoder.
• Separates the excitatation and vocal tract parameters.
• Cepstrum coefficients are obtained from the inverse Fourier Transform of the log magnitude
value of soectrum produced.
• Low frequency coefficients form the vocal tract envelope
• High frequency coefficients form the excitation parameters that foem a periodic pulse train.
• Linear filtering is done to separate the coefficients.
• At receiver coefficients are fourier transformed and vocal tract impulse response is formed.
• Convolving the impulse response with synthetic excitation signal(random noise or pulse
train) , original speech is reconstructed.
• This is a form of sub band coding.
VOICE EXCITED VOCODER
• Hybrid combination of PCM transmission for low frequency band and
channel vocoding of high frequency band.
• Designed for operation at 7.2-9.6kbps.
 LPC
• Algorithm
1) Speech sample is considered as linear combination of previous samples.
2) Speech is sampled, stored and analysed. Coefficients are calculated from samples.
3) Due to long term correlation from samples voiced and unvoiced signals are accurately catogerized.
• In this vocoder articulation tract is represented as recursive digital filter whose resonance (frequency
response) gives a set of filter coefficients.
• Computation of these coefficients based on mathematical optimization procedure results in LPC model.
• Widely used in speech telephony
• Advantages:
1) Manipulation facilities
2) Narrow analogy
3) Pitch & articulation tract parameters obtained by LPC coefficients that are directly accessible, audible
voice characteristics are highly influenced.
Regular Pulse Excited LPC

Vocoder parameter extraction

• Analyses the signal as voiced and unvoiced.
• After determining the period of voiced sounds, periodicity is encoded
& coefficients are transmitted.
• For the transition from voiced signal to unvoiced signal, a code is
transmitted that stops the receiver from generating periodic pulses &
makes it generate random pulses that is noise like nature.
Adaptive Systems for CELP
(a) Forward (b) Backward
General, Simplified CELP Decoder
• Used in telephony coding
• Coder is optimized to find the best match of the signal by using
codebook obtained through vector quantization method.
 Mixed Excitation LPC

• Four additional features:

1. Aperiodic pulses:
i) Vocoder synthesizes speech using periodic or aperiodic pulses.
ii) Aperiodic pulses usually used during transition between voiced and unvoiced segment
2. Pulse dispersion: fixed pulse dispersion filter is used
3. Mixed excitation :
i) multiband mixing
ii) stimulates frequency dependent voicing strength using adaptive filtering using structure
based on fixed filter bank.
• Adaptive spectral enhancement:
This filter is based on poles of the LPC vocal tract filter and used to enhance the formant structure in
synthetic speech.
Improvesmatch btwn synthetic & natural bandpass wave forms and introduces more natural quality to the
speech output.
 QCELP vocoder
• The basic CELP algorithm is one of the AbS methods widely used in
the low bit rate of speech coding.

• Qualcomm code excited linear prediction (QCELP) is also a CELP

algorithm, but it differs from the traditional CELP in that it
dynamically adjusts the encoded data rate based on speech signal
energy, background noise, and other speech characteristics.
7.9 SPREAD SPECTRUM MODULATION
It is a wideband modulation Technique.
Spread spectrum transmission offers the following three main
advantages over fixed frequency transmission:
(a) Spread spectrum signals are highly resistant to noise and
interference. The process of re-collecting a spread signal
spreads out noise and interference, causing them to recede
into the background.
(b) Spread spectrum signals are difficult to intercept.
(c) Spread spectrum transmissions can share a frequency band
with many types of conventional transmissions with minimal
interference. These signals add minimal noise to the narrow-
frequency communications, and vice versa. As a result,
bandwidth can be utilized more efficiently.
7.10 PSEUDO-NOISE
CODES, PROPERTIES,
AND CODE GENERATION

Time and frequency domain appearances

A unique code is used to

spread and despread the
signal using the logic
shown in Figure (a). This
unique code is the DS,
known as the pseudo-
noise (PN) code.
PN Code, Bit Rate and Spectrum
• Assume that the clock rate is provided for generating
each bit of a PN code (with one clock cycle, one bit of
PN code comes out).
• The bit rate of a PN code is called the chip rate (1/
tchip ), which is 10 times or more than the data bit
rate.
• The smallest time increment in the sequences of
certain period or duration is t chip and is known as a
time chip . The total period consists of Nc time chips.
• The chip rate decides the final transmission spectrum
of the DSSS system.
Autocorrelation Property of DSSS

Auto and Cross-correlation

Equations

The autocorrelation should be maximal

for the DS or PN codes so that correct PN
signal can be identified at the receiver
from the numerous coexisting signals.
Properties of Pseudo-noise Codes
• Balance property
• Run length property
• Autocorrelation property
In General for PN Code
1. In every period, the number of +1’s differs from that of −1’s by
exactly one (balance property). Hence, Nc is an odd number.
2. In every period, half of the runs of the same sign have length one,
one-fourth have length two, one-eighth have length three, and so
forth. In addition, the number of positive runs equals that of
negative runs (run property).
3. The autocorrelation of a periodic sequence is two valued, that is,
N c for shifts 0, N c , 2 N c , 3 N c , and so on and −1 otherwise
(without normalization).
7.11 DIRECT SEQUENCE SPREAD SPECTRUM
SYSTEM

Simplified diagram for biphase modulation

Transmitter Process
Receiver Process

Carrier demodulation and despreading of SSM signal to get original data

Detection of signal and despreading operations can be

either by
active method
or
passive method
Waveforms for Example 7.5
Spectral Density, Bandwidth, and Processing Gain

-----Spectral density of message signal

Processing gain
Spectral density of binary PN sequence

For Bi-phase Modulation

For Quadriphase Modulation

Rake Receiver Configurations

(a) With five fingers, one

combiner, and one
integrator

(b) Reception with multiple

integrators at rake
receiver and processing
DSSS System Performance
• Performance Parameters
• Interference Rejection

• Antijam characteristics

• Energy and Bandwidth Efficiency

Narrowband and Wideband Interference
Rejection
Near Far Problem
Power Control

(a) Open loop (b) Closed loop

7.12 FREQUENCY HOPPING SPREAD
SPECTRUM— TRANSMITTER AND RECEIVER

Two schemes---Slow Frequency Hopping and Fast Frequency Hopping

FHSS Generator Diagram
Non-Coherent FHSS Receiver
Data and Time-frequency Plane for
Example 7.10
 7.13 TIME HOPPING SPREAD SPECTRUM

TH (a) Concept (b) Waveforms showing THSS signal formation on bit-by-bit basis (c) TH with
variable time slots (bit by bit)
• For FHSS

• For THSS
Comparison of SSM Methods
• For FHSS

• For THSS
Comparison of SSM Methods
7.14 HYBRID SPREAD SPECTRUM SYSTEMS
• The use of hybrid techniques attempt to capitalize upon the advantages of
a particular method while avoiding the disadvantages.
• DS, suffers heavily from the near–far effect, which makes this technique
hard to apply to systems without the ability of power control, but its
implementation is inexpensive.
• The PN code generators are easy to implement and the spreading
operation itself can be simply performed by XOR ports.
• FH effectively suppresses the near–far effect and reduces the need for
power control. However, implementation of the (fast) hopping frequency
synthesizer required for a reasonable spreading gain is more problematic
in terms of higher silicon cost and increased power consumption.
• Selection of SFH/FFH also has its own pros and cones.
• Solutions are
PN/FH, PN/TH, FH/TH, and PN/FH/TH.
7.15 MULTICARRIER MODULATION TECHNIQUES
• Basic Principles of Orthogonality
Two periodic signals are orthogonal when the integral of their
product over one period is equal to zero and they have an
integral number of cycles in the fundamental period; For N =
period of k samples
Subcarrier Setting—Conceptual Representation

Frequency-to-time domain
conversion (a) Orthogonal
subcarriers setting in the
frequency domain with 32 point
IFFT bin (b) Corresponding time
domain interpretations for
interval N = 32 samples (c) Plot of
four subcarriers on the same
time axis for the addition of
subcarriers in the time domain to
get the OFDM baseband
Example of Four Subcarriers with Three Symbols per
Subcarrier
at OFDM Modulator Stage

(a) Lattice representation of carrier assignment planning

Cont’d on Next slide

(b) Concept of carrier frequency allocation to the symbols (a typical case in which three
symbols are assigned to a carrier), serial-to-parallel conversion, and mapping into frequency
domain components after modulation
 Four subcarriers making the total occupied bandwidth for an OFDM baseband signal
(a) Modulated subcarriers with spectral setting (b) Overall bandwidth

Orthogonal signals in the frequency domain:

peak of one signal occurs at the null of its nearer subcarrier
FDM v/s OFDM

Comparing
Spectral
Efficiency
(a) Spectrum saving
due to multicarrier
modulation

(b) Better spectral

efficiency of OFDM
compared to other
techniques
7.16 ORTHOGONAL FREQUENCY DIVISION MULTIPLEXING
TRANSMITTER AND RECEIVER

• Three main stages of processing at transmitter

• Channel coding stage
• Modulation stage
• Upconversion stage

• At receiver exactly reverse procedures are applied

including channel correction processes

• Details of Transmitter and Receiver are given in the next

slide.
 OFDM Transmitter and Receiver

Transmitter Stage Processing

Fading effect on OFDM received signal

Important Signal Processing Stages
• Serial to parallel conversion
• Symbol mapping
• Modulation of data
• IFFT—frequency –to-time domain conversion
Frequency to time domain conversion and IFFT
acting as a summer
Advantage of Cyclic Prefix Addition

Removal of muti path effect

Method of Including Cyclic Prefix
Radio Frequency Upconversion Stage
• Analog or digital techniques but final form of transmission is analog in
nature
RF Upconversion
Frequency Control Required in OFDM
Demodulation Steps in OFDM

• Partition the input stream into vectors representing each symbol

period.
• Take the FFT of each symbol period vector.
• Extract the carrier FFT bins and calculate the phase of each bin.
• Calculate the phase difference, from one symbol period to the next,
for each carrier.
• Decode each phase into binary data.
• Sort the data in the appropriate order.
Effect of the number of Subcarriers and Guard
time duration on the System Performance

• For a given number of subcarriers, increasing the

guard time duration reduces the ISI due to the
decrease in the delay spread relative to the symbol
time, but it also reduces the power efficiency and
bandwidth efficiency.
• For a given signal bandwidth, increasing the number
of subcarriers increases the power efficiency, but it
also increases the symbol duration and results in a
system more sensitive to Doppler spread.
Applications of OFDM
• WiFi
• WiMAX
• Ultra-wideband
• Digital audio broadcasting and digital video broadcasting
• Long-term evolution
 8.1 Zero Inter symbol
Interference Communication
Techniques
Intersymbol interference (ISI) is a form of distortion in a signal. Here,
a received symbol interferes with the subsequent symbols, which
makes the communication less reliable. ISI is caused due to
multipath propagation or the inherent non-linear frequency
response of a channel.

ISI between pulses or symbols (ideally square shape)

 Nyquist Criteria for Zero Inter
symbol Interference
• Nyquist showed that the theoretical minimum bandwidth needed to
detect Rs symbols without ISI is Rs /2 Hz.
• The sinc-shaped (t/T) pulse is called the ideal Nyquist pulse ; its
multiple lobes comprise a main lobe and side lobes that are infinitely
long.
• Nyquist established that if each pulse of a received sequence is of the
form sinc (t/T), the pulses can be detected without ISI.
 Nyquist Pulse
There are two successive pulses s ( t ) and s ( t − T ). Even though
s ( t ) has long tails, the figure shows a tail passing through zero
amplitude at the instant t = T when s ( t − T ) is to be sampled;
likewise, all tails pass through zero amplitude when any other
pulse of the sequence is to be sampled. They satisfy the
orthogonality condition.
 Bandwidth Constraint and
Nyquist Filter
• If we operate the system at smaller bandwidths, then according to the
Nyquist condition, the pulse would spread in time, which would degrade
the system bit error rate (BER) performance due to increased ISI.

• A prudent goal is to compress the bandwidth of the data impulses to a

reasonably small bandwidth greater than the Nyquist minima. This is
accomplished with a Nyquist filter. Without such a measure, each pulse
extends into every other pulse in the entire sequence.

• The terms Nyquist filter and Nyquist pulse are often used to describe the
general class of filtering and pulse shaping that satisfies zero ISI at the
sampling points. Transversal Filter, raised cosine, square root raised cosine,
Gaussian, Chebyshev etc are the filters for pulse shaping to reduce ISI.
 Filtering (Pulse Shaping)
• There are various filters throughout the system—in the
transmitter, channel, and receiver. Taking all these filtering
effects into one overall equivalent system transfer
function,
• where Ht(f) characterizes the transmitting filter shaping
the pulse, Hc(f) is the filtering within the channel, and Hr(f)
is the receiving or equalizing filter. System shown in the
figure below.
 Raised Cosine Filter
• Time domain response of raised cosine filter

The sharpness of a raised cosine filter is described by the roll-off

factor (α), which decides the slope of the low-pass filter cut-off
region. The value of α gives a direct measure of the occupied
bandwidth of the system and is calculated as
Raised Cosine Filter Response
for Different Values
An α of zero is impossible to implement. α is sometimes called the excess
bandwidth factor , as it indicates the amount of occupied bandwidth that will be
required in excess of the ideal occupied bandwidth (which would be the same
as the symbol rate).
Raised Cosine Filter Response
for Different Values
QPSK vector diagrams—effect of  (can be
observed by simulation)
(a) Without filtering (b)  = 0.75
(c)  = 0.375
Square Root Raised Cosine
Filter

With premodulation and predemodulation filtering---The Combined response is that of a

Nyquist filter
 Gaussian Pulse Shaping
Filter
Gaussian filter is normally used in Gaussian The Gaussian filter has a Gaussian
minimum shift keying (GMSK) and Gaussian shape in both the time and
frequency shift keying (GFSK) modulation
schemes. frequency domains, and it does not
ring like the raised cosine filters.
Its effects in the time domain are
relatively short and each symbol
interacts significantly (or causes ISI)
with only the preceding and
succeeding symbols. This makes
amplifiers easier to build and more
efficient. Important parameter
bandwidth-time (BT) product.

Time Response
 Chebyshev Equiripple Finite
Impulse Response Filter
A Chebyshev equiripple finite
impulse response (FIR) filter is
used for baseband filtering in
IS-95 code division multiple
access (CDMA).

Reduction of leakage to
adjacent radio frequency (RF)
channels is accomplished by
using a filter with a very
sharp shape factor using an α
value of only 0.113.
Two types of Chebyshev low-pass filters, both are based on Chebyshev polynomials.
(a) The type I filter has an all-pole transfer function, and it has an equiripple passband and a
monotonically decreasing stopband.
(b) A type II low-pass filter has both poles and zeros. Its passband is monotonically decreasing
and it has an equiripple stopband.
 Window Techniques
• Windowing suppresses discontinuities and avoids the
broadening of the frequency spectrum caused by
discontinuities.
 Window Techniques
Windowing can narrow the spectrum, but it is important to remember that windowing
is really a distortion of the original signal. It adds BER in the performance but improves
spectral efficiencies at the same time. Using the windowing function in a system is a
compromise.

The result of applying a window function without proper thought

Examples of the Windowing
Functions
(in discrete form)
 8.2 Detection Strategies
• The expected signal must be detected from among the
various coexisting signals at the receiving end.
• Detection theory is well established and is a means to
quantify the ability to differentiate between information-
bearing energy patterns and random energy patterns (such
as noise) that distract from the information.
• The basis for the signal detection theory is that nearly all
reasoning and decision-making takes place in the presence of
some uncertainty. Mostly, threshold based detection
method.
• Graphical interpretations, receiver operating characteristics,
and discriminability index are used for decision-making.
 8.2 Detection Strategies
• Suppose there are two states (hypotheses) H 0 and H 1
at detection level hypothetically, only two states are
possible—the presence of the required signal within
noise or only noise with no desired signal;
mathematically, this can be represented as follows with
sent, received and noise(AWGN) samples:

• Detection is based on some function T of the received

samples, which is compared to a threshold γ. The
example is given in the next slide.
 8.2 Detection Strategies
•The probability of a false alarm
PFA is the probability that H1 is
selected even when H0 is
actually true; that is, PFA = P ( T (
r ) > γ, H0).
•The probability of miss PM is
the probability that H0 is
selected when H1 is true. •The optimal detectors use the maximum
likelihood test in which probability density
•The probability of detection is functions (PDFs) are used if the parameters
PD = 1 − PM and is the are unknown and random but their PDFs are
known.
probability that H1 is selected •In some applications, it is possible to assign
when it is actually true; that is, prior probabilities to the possible
PD = P ( T ( r ) > γ; H1). hypothesis.
•Energy detectors are used sometimes.
 8.3 Matched Filter
A matched filter correlates the incoming signal with a locally stored reference
copy of the transmit waveform. The matched filter maximizes the SNR for a
known signal. It can be an optimal detector under the following conditions
• The channel produces AWGN.
• The channel is LTI.
• Exact time reference is available (the signal amplitude as a function of time is
precisely known). A possible matched filter receiver is shown below.
 8.4 Diversity Techniques
Diversity techniques are used for improving the reliability of a message signal by
utilizing two or more communication channels with different characteristics.
Multiple versions of the same signal may be transmitted and/or received and
combined in the receiver.
In the figure, different independent fading paths are shown with different CIRs h1(t)
to hn(t), and the distortion in the signal is observed due to addition of interfering
signals I1 to In and white Gaussian noise. The strongest signal is then picked from the
received signals. This signal can be further equalized and demodulated to receive
the digital signal with minimum BER.
 8.4 Diversity Techniques
• Diversity combining is the process to extract the main
transmitted signal with minimum channel effects out of many
versions of a signal received.
• Depending upon the type of fading, diversity techniques have
been categorized as
• micro-diversity techniques
• macro-diversity techniques
• Micro-diversity technique--These techniques are used in a
small-scale fading environment--two antennas are separated
by a fraction of a metre.
• Macro-diversity technique--These techniques are used in a
large-scale fading environment—antennas are quite far apart
and not shadowed.
 8.4 Diversity Techniques
• Space diversity--microscopic or macroscopic diversity technique; Space
separation of half of the wavelength is sufficient to obtain two uncorrelated
signals.
• Polarization diversity--obstacles scatter waves differently depending on their
polarization. Antennas can transmit either a horizontal or a vertical polarized
wave. When both waves are transmitted simultaneously, received signals will
exhibit uncorrelated fading statistics.
• Angle diversity--the received signals arrive at the antenna via several paths,
each with a different angle of arrival, the signal component can be isolated
by using directional antennas.
• Frequency diversity—information is transmitted on more than one carrier
frequency, because the frequencies separated by more than the coherence
bandwidth of the channel will not experience the same fading.
• Time diversity--multiple versions of the same signal are transmitted at
different time instants.
• Joint diversity—combination of above mentioned techniques
 8.5 Diversity Combining
Techniques
• Space diversity reception or combining methods can be classified into
four types:
(a) Selection combining
(b) Threshold combining
(c) Maximal ratio combining
(d) Equal gain combining
 Selection Combining
Simplest Technique--It is based on the principle of selecting the
best signal (the largest energy or SNR) among all the signals
received from different branches. Scheme can be represented as
follows:
 Threshold Combining
Threshold combining is a special form of selection diversity that is less
expensive to implement. Here, a limited number of signals are considered
for the extraction purpose using a threshold level. Scanning and feedback
mechanisms are used here.
 Equal Gain Combining
Equal gain combining (EGC) is better than selection diversity and is almost as
good as maximum ratio combining (MRC), but is less complex in terms of the
signal processing and feedback part. The idea behind this technique will be
clear from Figure.
Maximum Ratio Combining
•For noise-limited systems without interference, MRC results in the best SNR.
•Here, all the incoming signals from all the M branches are weighted according to
their individual signal voltage to noise power ratios and then added.
•All the individual signals must be co-phased before being added. This requires an
individual receiver circuitry and a phasing circuit for each antenna element.
•It produces an acceptable SNR compared to other techniques.
• It has extremely good signal processing and complex hardware.
 8.6 Introduction to Multiple
Input, Multiple Output
Three basic link performance Systems
parameters completely describe the quality of any
wireless link: speed (or spectrum), range (or coverage), and reliability (or security).

MIMO offers greater spectral efficiency as

compared to SISO, SIMO, and MISO systems.
Higher data rates, greater range, increased
number of users, and enhanced reliability.

•It exploits the use of multiple signals (space

diversity) into the wireless medium and multiple
signals received from the wireless medium to
improve the wireless channel performance.
• It can provide a combination of a multi-antenna
system with a multicarrier system.
 Spatial Diversity
• Presently, four different types of multi-antenna systems
can be categorized based on diversity (input and output
refer to the number of antennas):

(a) Single input, single output (SISO)—no diversity

(b) Single input, multiple outputs (SIMO)— receive diversity
(c) Multiple inputs, single output (MISO)— transmit
diversity
(d) Multiple inputs, multiple outputs (MIMO)—transmit–
receive diversity
 Example of SIMO

In a SIMO channel, the concept of MRC is offered as a way to exploit the receive
diversity. The error probability achieved by MRC is to be much smaller than that
corresponding to a SISO channel.
To perform MRC, the receiver has to know the fading, or, in other words, the
receiver has to have access to the CSI. This is usually done by sending some
known signal through the channel.
 Comparison of Channel
Capacities of Various Multi-
antenna Systems
• According to Shannon, the limit on the channel capacity
is given by (for SISO system)

• For SIMO system (M receiving antennas)

• For MISO system (N transmitting antennas)

• For MIMO system (N transmitting and M receiving

antennas)
 Smart Antenna
• Smart antenna systems used with spatial diversity exploit the concept of MIMO
intelligently.
• A smart antenna can automatically change the directionality of its radiation
patterns in response to its signal environment.
• Terms that are commonly associated with various aspects of smart antenna
include phased array, space division multiple access (SDMA), spatial processing,
digital beamforming, coherent combining, and adaptive antenna systems.
• Smart antenna systems fall into two main categories: (a) switched-beam
systems and (b) adaptive array systems. Both the systems direct a main lobe (or
radio beam) towards individual users and attempt to reject interference or
noise from the outside of that main lobe.
• In smart antennas, the data is transmitted over a vector channel.
• Normally, a smart antenna system performs better in LOS or close-to-LOS
systems.
 Spatial Multiplexing in MIMO
The underlying mathematical nature of spatial multiplexed MIMO,
where data is transmitted over a matrix rather than a vector channel,
creates new and enormous opportunities beyond just the added
diversity or array gain benefits— the spectrum efficiency.
Spatially Multiplexed MIMO
System
3 x 3 MIMO Channel
Modeling
It is common to model a wireless channel
as a sum of two components—
an LOS component
an NLOS component.
The MIMO channel model is a matrix in
mathematical form. To analyse the
receptions through it, we have to use
eigenvectors and singular-value
decomposition techniques.

The Rician factor is the ratio between the power of the LOS
component and the mean power of the NLOS component.
In outdoor environment LOS dominates while in indoor
environment NLOS. H is the complex channel matrix.
 8.7 Channel Estimation
Techniques
• In the detection part, we have two options to inverse the distortion due to channel.
We can do channel estimation followed by channel inversion or two separate tasks—
or we can do channel equalization directly based on certain criterion such as
minimum mean square error (MMSE).

• Channel estimation is an autoregressive process that may be performed with a

number of iterations.

• An autoregressive model specifies that the output variable depends linearly on its
own previous values or some known values.

• Channel estimation is the estimation of the channel Impulse Response (CIR) at the
receiver.

• Three techniques of CE—

• CSI based ---pilot or training sequence based
• Blind--- statistical data or prior knowledge based
• Semi-blind
 Channel Impulse Response
Estimation Basic
• In general, channel processes are wide-sense stationary.
• Conceptually, on a stationary wireless channel, if an impulse is
transmitted, then multiple delayed versions of impulses will be
received at the receiver at different instants of time.
• These impulses are non-correlated and with reducing amplitudes
with time. They are just like delayed samples, and hence, a
differential equation with coefficient values can be correlated with
this concept. FIR filter model is adopted.
• The impulse response he(n) is time varying and can be represented
as

• Next state estimation is based on

Comparison between CE
Techniques
8.8 Equalization Techniques
Basic concept of Equalization in mathematical form and Block diagram

An equalizer is an inverse channel filter to mitigate the

unpredicted channel problems.
8.8 Equalization Techniques

Channel equalization concept in a digital

communication system
using matched filter and decision-making
Transversal Filter
Multiplierless Transversal
Filter
 Adaptive Equalizers

Adaptive equalizers can control the tap weight on the basis of

estimated channel coefficients from the received known
training sequences.
Decision Directed Feedback
Equalizer
8.9 Least Squares and Least
Mean Squares Algorithms
LS--Least squares (LS) means that the overall solution minimizes
the sum of the squares of the errors made in the results of every
single equation.

The LS and LMS algorithms represent the regression processes to

adapt the system continuously.

Least mean squares (LMS) algorithms are a class of adaptive filters

used to mimic a desired filter by finding the filter coefficients that
relate to producing the LMS of the error signal (difference between
the desired and actual signals). It is a stochastic gradient descent
method in which the filter is adapted based on the error at the
current time.
LMS Formulation Diagram