You are on page 1of 63

Chapter 9.

Clutter and Moving Target


Indicator (MTI)

Hankuk Aviation Univ.

RSP Lab

Ground Radar - Environment

Hankuk Aviation Univ.

RSP Lab

Radar Environments

< Radar signals and Interference>

Hankuk Aviation Univ.

Noise :
in the receiver, ant, tx line
outside from sun random.
Random motion at all temp.
above absolute zero.

Clutter : Unwanted signal


echo from sea, land, weather

ECM :
electromagnetic
countermeasures noise jamming

EMI :
friendly sources such as other
radar, comm. sys, friendly jammer

Spillover : internal clutter

RSP Lab

9.1 Clutter Definition


Clutter Definition : Clutter is unwanted radar returns
that may interfere with normal radar operations.
Type : Mainloabe Clutter & Sidelobe Clutter
1) Surface clutter :
- Ground clutter : trees, vegetation, ground terrain, man-made structure
- Sea clutter : sea surface (sea clutter)
2) Volume clutter : chaff, rain, birds, insects

Notes: Individual clutter components : random phase and amplitude

Clutter signal level >> receiver noise level


Radars ability to detect targets signal-to-clutter ratio (SCR)
Hankuk Aviation Univ.

RSP Lab

Radar Clutter Type


Clutter

AREA
Land
-mountains
-woods
-vegetated
farmland
-desert

VOLUME

OTHER

Weather
- rain
- snow

Moving vehicles

Chaff

Insects

Dust storm

Angles

Birds

SEA
Hankuk Aviation Univ.

RSP Lab

Signal-to-Clutter Ratio (SCR)


Average Clutter RCS is

c 0 Ac

(9.1)

where 0 clutter scattering coefficien t (m 2 / m 2 )


Ac clutter area

-Propagation factor :

- constructive/destructive interference of the electromagnetic waves diffracted


from an object (target or clutter)
-Target/clutter returns with different angles of arrival of different
propagation factors

t Ft 2 Fr2
SCR
c Fc2

(9.2)

where Fc clutter propagation factor


Ft / Fr Tx / Rx propagation factors for target. in many case, Ft Fr
Hankuk Aviation Univ.

RSP Lab

9.2 Surface Clutter


- Surface clutter includes both land and sea clutter.
- Area clutter is concern for
1) Airborne radars in the look-down mode
2) ground-based radars when searching for targets at low grazing angle.
- Grazing angle ( g ) : angle from the surface of the earth to the main axis of
the illuminating beam.

< Definition of grazing angle >


Hankuk Aviation Univ.

RSP Lab

Grazing Angle vs. Scattering Coefficient


- Three factors affect the amount of clutter in the radar beam.
1) Grazing angle 2) Surface roughness 3) Radar wavelength
- Smaller wavelength larger scattering coefficient

0
< Dependency of

Hankuk Aviation Univ.

on the grazing angle >


RSP Lab

Critical Grazing Angle


- Low grazing angle region : from zero to the critical angle.

- Critical angle : angle below which a surface is considered to be smooth, and


above which a surface is considered to be rough.
- hrms = rms of a surface height irregularity

- According to the Rayleigh criteria the surface is considered to be smooth if

4 hrms

sin g

(9.3)

- Due to surface height irregularity, the rough path is longer than the

smooth path
by a distance 2hrms sin. g
Hankuk Aviation Univ.

RSP Lab

Rough Surface
- This path difference translates into a phase differential

2hrms sin g

(9.4)

< Rough surface definition >

- When (first null), Grazing angle g = critical angle gc


Hankuk Aviation Univ.

RSP Lab

Rough Surface
4 hrms

sin gc

(9.5)

or equivalently,

gc sin 1

hrms

(9.6)

- In the case of sea clutter, the rms surface height irregularity is


1.72
hrms 0.025 0.046 Sstate

(9.7)

- Clutter at low grazing angle diffuse clutter : large number of clutter


returns in the radar beam (non-coherent
reflections)
- Clutter in the high grazing angle region is more specular (coherent
reflections)
Hankuk Aviation Univ.

RSP Lab

Radar Equation for Area Clutter


- Airborne radar in the look-down mode case.

Elliptical shape

- Footprint size = f ( g ,3dB )


- Footprint is divided into many ground range bins each size (c / 2) sec g
Hankuk Aviation Univ.

RSP Lab

Ground Clutter Model



- 0

-
( 3% )

(A)

A 2R(c/2)tan(az / 2) sec( )
: 3

Pt G 2 2 Pt G 2 0 A2
Pc

3 4
(4 ) R
(4 )3 R 4
Hankuk Aviation Univ.

RSP Lab

Ground Clutter Geometry


Radar

az Half Power azimuth beamwidth

2 R tan(el / 2)
tan
c / 2

2R
tan(az / 2)

(c / 2 ) sec

A 2 R(c / 2 ) tan(az / 2) sec


Hankuk Aviation Univ.

where

2R tan(el / 2)
tan
c / 2
RSP Lab

Radar Equation for Area Clutter


- Clutter area Ac is
Ac R 3dB

c
sec g
2

(9.8)

- Power received by the radar from


a scatterer within Ac is
Pt G 22 t
St
(4 )3 R 4
where t target RCS

(9.9)

- Received power from clutter is


< Footprint definition >

- SCR for area clutter is


( SCR) A c
Hankuk Aviation Univ.

Pt G 22 c
S Ac
(4 )3 R 4
2 t cos g

03dB Rc

(9.10)

(9.11)
RSP Lab

Example 9.1
Example 9.1 : Consider an airborne radar. Let the antenna 3dB beam width
be 3dB 0.02rad , the pulse width 2s, range R 20km, and grazing angle

g 20o. Assume target RCS t 1m 2 , and clutter reflection coefficien t


o 0.0136 m 2 / m 2 . Compute the SCR.
( SCR) Ac

2 t cos g

0 3dB Rc

(2)(1)(cos 20)
4

2
.
48

10
(0.0136)(0.02)(20000)(3 108 )(2 10 6 )

It follows that
( SCR ) Ac 36.06dB
thus, for relible detection the radar must somehow increase its SCR by
at least (36 X )dB, where X is on the order of 10dB or better .

Hankuk Aviation Univ.

RSP Lab

Sea Clutter Model



- Sea State
-

M v 0.8 W
v 0.105 W

(A)

A 2R(c/2)tan( az / 2) sec( )
R

Pc

Pt G 2 2 Pt G 2 0 A2
Pc

3 4
(4 ) R
(4 )3 R 4
Hankuk Aviation Univ.

RSP Lab

9.3 Volume clutter


- Volume clutter includes rain, chaff, birds, insects the volume clutter
coefficient is expressed in squared meters

- Birds, insects and other flying particles are referred to as angel clutter
the average RCS as a function of the weight of the bird or insect is reported
as,

b dBsm 46 5.8 log wb

(9.12)

where, wb is the individual weight in grams


- Bird and insect RCS are also function of frequency

ex) pigeons RCS is -26dBsm at S-band, -27dBsm at X-band


Hankuk Aviation Univ.

RSP Lab

Volume Clutter Model

- Wind Shear, Beam Broadening, Turbulence


-
Az/El/(V)

V R 2 el az / 4 c / 2

Pc
Pt G 2 2 Pt G 2 vV2
Pc

3 4
(4 ) R
(4 )3 R 4

c / 2
Rel

Hankuk Aviation Univ.

To radar

Raz

R 2 el az c
V

4

RSP Lab

Chaff RCS
- Chaff is used as ECM technique by hostile forces.
It consist of a large number of dipole reflectors (large RCS values).
Maximum chaff RCS occurs when dipole length L is one half radar wavelength.
Average RCS for single dipole when broadside is,

chaff1 0.882

(9.13)

and for an average aspect angle, it drops to

chaff1 0.152

(9.14)

where, the subscript chaff1 indicate a single dipole


- The total chaff RCS within radar resolution volume is,

chaff 0.152 N D

(9.15)

where, ND is total number of dipoles in the resolution volume


Hankuk Aviation Univ.

RSP Lab

Rain Clutter Model


M v V0 cos

v shear turb beam


2

shear 0.42kRel
turb 1.0

(m / sec)

fall 1.0 sin( )

2
fall

k wind shear gradient (4m /(sec)( Km))


R slant range( Km)

(m / sec)

beam 0.42V0az sin

(m / sec)

(m / sec)

el elevation beamwidth(radians)
az azimuth beamwidth(radians)
azimuth relative to wind direction at beam center
elevation angle
V0 wind speed at beam center(m/s ec)

Hankuk Aviation Univ.

RSP Lab

Weather and rain clutter


- Weather or rain clutter is easier to suppress than chaff, since rain can be as
perfect small spheres.

- We can use the Rayleigh approximation of perfect sphere to rain droplets


RCS Rayleigh approximation is given as,

9r kr
2

(9.16)

where, k 2 / and r is radius of a rain droplet


- Electromagnetic wave when reflected from perfect sphere become strongly
co-polarized (same polarization as incident waves) .

Therefore backscatter energy from rain retains the same polarization as


incident waves, but reversed direction of propagation.
So, radar suppress rain clutter by co-polarizing the radar antenna.
Hankuk Aviation Univ.

RSP Lab

Resolution volume
- Defining as RCS per unit resolution volume VW, it computed as,
N

i
i 1

(9.17)

where, N is the total number of scatterers within the resolution volume


- Total RCS of a single resolution volume is,
N

W i vW

(9.18)

i 1

- A resolution volume is in Fig 9.6, and is approximated by

VW

a e R 2c

(9.19)
8
where, a ,b are antenna beam width in az, el, is pulse width, R is range
Hankuk Aviation Univ.

RSP Lab

Weather clutter coefficient

<Fig 9.6 Definition of a resolution volume>

- Consider a propagation medium with an index of refraction m.


The ith rain droplet RCS approximation in this medium is,
where,

5 2 6
i 4 K Di

m2 1
2
K 2
m 2

(9.20)

(9.21)

where, Di is the ith droplet diameter


Hankuk Aviation Univ.

RSP Lab

Weather clutter coefficient


- For example, temperatures between 32F and 68F yield,

5 6
i 0.93 4 Di

- and for ice (9.20) can be approximated by,


5 6
i 0.2 4 Di

- Substituting (9.20) into (9.17) yields


5 2
4 K Z

where the weather clutter coefficient Z is defined as


N

Z Di

(9.22)

(9.23)

(9.24)

(9.25)

i 1

- In general, the units of Z are often expressed in millimeter6/m3


Hankuk Aviation Univ.

RSP Lab

Radar equation for volume clutter


- The total power received by radar from t at R is
Pt G 2 2 t
St
(4 )3 R 4
- The weather clutter power received by the radar is

(9.26)

Pt G 22 W
SW
(9.27)
3 4
(4 ) R
- Using (9.18) and (9.19) into (9.27) and collecting terms yield,
N
Pt G 22 2
SW
R a e c i
(9.28)
3 4
(4 ) R 8
i 1
- SCR for weather clutter is computed by dividing (9.26) by (9.28), more
precisely,
8 t
SCR V St
N
(9.29)
SW
2
a e cR i
where V is used to denote volume clutter.
Hankuk Aviation Univ.

i 1

RSP Lab

9.4 Clutter statistical models


- Clutter is statistically described by a probability distribution function.

The type of distribution depends on the nature of clutter itself (sea, land,
volume), radar operating frequency and the grazing angle.
- If probability of receiving scatterer is statistically independent of another
scatterer, then, the clutter may be modeled using a Rayleigh distribution,
x2
2x
; x 0
f x
exp
x0
x0

(9.30)

where x0 is the mean squared value of x


- The log-nomal distribution best describes land clutter at low grazing angles.

it also fits sea clutter in the plateau region. It given by,

Hankuk Aviation Univ.

RSP Lab

Clutter statistical models


ln x ln xm 2
1
;
f x
exp
2

2
2

x0

(9.31)

where xm is the median of the random variable x, is the standard deviation


of ln(x).
- Weibull distribution is used to model clutter at low grazing angles for 1 to 10
Ghz. Weibull probability density function is determined by the Weibull slope

parameter and a median scatter coefficient 0 and given by,

xb
f x
exp ; x 0
0
0
where, b=1/a is known as the shape prameter.
bx b 1

(9.32)

when b=2 the Weibull distribution becomes a Rayleigh distribution.


Hankuk Aviation Univ.

RSP Lab

9.5 Clutter Spectrum


- Clutter is not always stationary : wind speed, motion of the radar scanning
antenna Doppler frequency spread
- In Ground Radar
clutter spectrum : concentrated around f 0 and integer multiples of the radar
PRF f r
: some small spreading

- clutter power spectrum : fixed (stationary) + random (frequency spreading)


for most cases, Gaussian

Hankuk Aviation Univ.

RSP Lab

Clutter Spectrum
- denote the fixed to the random power ratio by

W2

clutter spectrum
0 2
W2
0

0
Sc w 0
exp
2
2
2
2
1W
1 W 2 2

stationary clutter

(9.33)

frequency spreading

0 2f o , : rms frequency spread component, 0 : Weibull parameter


- Clutter power :
concentrated around zero Doppler with some spreading
(typically less than 100Hz)

Hankuk Aviation Univ.

RSP Lab

Ground Clutter - Environment

Hankuk Aviation Univ.

RSP Lab

Clutter Radial Velocity Characteristics

Hankuk Aviation Univ.

RSP Lab

Clutter PSD
- Model clutter using a Gaussian-shaped power spectrum
Sc w

Pc
2

0 2

exp
2
2

(9.34)

Pc : total clutter, : deviation, 0 : mean


Concentrated around DC and
integer multiples PRF

<Typical radar return PSD when clutter and target are present >

Hankuk Aviation Univ.

RSP Lab

Clutter Spectrum Characteristics

* Responce of a double canceller MTI to ground, rain , and chaff clitter

Hankuk Aviation Univ.

RSP Lab

9.6 Moving Target Indicator (MTI)


- In CW radar :
suppress clutter return by

ignoring the receiver output DC


- In Pulsed radar system:
suppress clutter return by using
special filter, MTI
- MTI filter :
deep stop-band at DC and
at integer multiples of the PRF

Hankuk Aviation Univ.

(a)
(b)
(c)

Typical radar return PSD


MTI filter frequency response
Output from an MTI filter
RSP Lab

Blind Speed
MTI Filter
- using delay line cancelers

- periodic frequency response (null at nf r )


Blind speed : target Doppler frequency=

vblind

nf r
2

nf r severely attenuate

; n0

(9.35)

- minimize the occurrence of blind speeds


PRF staggering : changing PRF between consecutive pulses
using high PRF

Hankuk Aviation Univ.

RSP Lab

Coherent MTI Radar Block Diagram

<Coherent MTI Radar Block Diagram>


Hankuk Aviation Univ.

RSP Lab

9.7 Single Delay Line Canceler


two-pulse canceler
ht : impulse response
y t xt * ht : output
delay T PRI

1
fr

< Single delay line canceler >

impulse response
Fourier transform
Z-domain

Hankuk Aviation Univ.

yt xt xt T

(9.36)

ht t t T

(9.37)

H 1 e jT
H z 1 z 1

2f

(9.38)
(9.39)

RSP Lab

Power gain (Single Delay line Canceler)


Power gain for the single delay line canceler response

H H H * 1 e jT 1 e jT
2

(9.40)

e jwt cos wt j sin wt

H 1 1 e jT e jT 21 cos T
2

(9.41)

2 2 cos 2 4sin 2

H 4sin T 2
2

Hankuk Aviation Univ.

(9.42)

RSP Lab

MATLAB Function single_canceler.m


single canceler period f r ( PRF ), peak f 2n 1 2 f r , nulls f nf r

< Single canceler frequency response >

Hankuk Aviation Univ.

RSP Lab

9.8 Double Delay Line Canceler


- Two basic configurations of a double delay line canceler
Double canceler are often called three-pulse canceler

< Two configurations for a double delay line canceler. >


Hankuk Aviation Univ.

RSP Lab

Double Delay Line Canceler


- The double line canceler impulse response is given by

h(t ) (t ) 2 (t T ) (t 2T )

(9.43)

- The power gain for the double delay line canceler is

H ( ) H1 ( ) H1 ( )
2

(9.44)

- It follows that

T
H ( ) 16 sin
2

(9.45)

- In the z-domain

H ( z ) (1 z 1 ) 2 1 2 z 1 z 2
Hankuk Aviation Univ.

(9.46)
RSP Lab

MATLAB Function double_canceler.m


- MATLAB Function double_canceler.m

[resp] double _ canceler ( fofr )

fofr is the number of periods desired.


Better response than the single canceler (deeper notch and flatter pass-band response)

< Normalized frequency response for single and double cancelers. >
Hankuk Aviation Univ.

RSP Lab

9.9 Delay Line with Feedback(Recursive Filters)


- The advantage of a recursive filter
shape the frequency response of the filter
- From the figure

y(t ) x(t ) (1 K )w(t ) (9.47)


v(t ) y(t ) w(t )

(9.48)

w(t ) v(t T )

(9.49)

< MTI recursive filter >

- Applying the z-transform to the above three equation yields

Y ( z) X ( z) (1 K )W ( z)

(9.50)

V ( z) Y ( z) W ( z)

(9.51)

W ( z ) z 1V ( z )

(9.52)

Hankuk Aviation Univ.

RSP Lab

Delay Line with Feedback(Recursive Filters)


- Solving for the transfer function

H ( z) Y ( z) / X ( z)

yields

1 z 1
H ( z)
1 Kz 1

(9.53)

- The modulus square of H (z ) is then equal to


(1 z 1 )(1 z )
2 ( z z 1 )
H ( z)

(1 Kz 1 )(1 Kz ) (1 K 2 ) K ( z z 1 )
2

(9.54)

- Using the transformation z e jT yields

z z 1 2 cos T

(9.55)

- Thus, Eq. (9.54) can now be rewritten as


2

H (e jT )

Hankuk Aviation Univ.

2(1 cos T )
(1 K 2 ) 2 K cos(T )

(9.56)

RSP Lab

Delay Line with Feedback(Recursive Filters)


- When K=0, Eq. (9.56) collapses to Eq. (9.42)

H ( ) 4sin T / 2
2

(9.42)

- By changing the gain factor K one can control of the filter response

< Frequency response corresponding to Eq.(9.56). >


Hankuk Aviation Univ.

RSP Lab

Delay Line with Feedback(Recursive Filters)


- In order to avoid oscillation due to the positive feedback
the value of K should be less than unity
- The value (1 K ) 1 is normally equal to the number of pulses received from
the target
ex) K=0.9 corresponds to ten pulses
K=0.98 corresponds to about fifty pulses

Hankuk Aviation Univ.

RSP Lab

9.10 PRF Staggering


- Blind speeds can pose serious limitations
performance of MTI radars
ability to perform adequate target detection
- Using PRF agility by changing the pulse repetition interval consecutive pulse
extend the first blind speed to tolerable values
- In order to show how PRF staggering
assume that two radars with distinct PRFs are utilized for detection
using two radars to alleviate the problem of blind speed is a very costly

option
- A more practical solution
to use a single radar with two or more different PRFs
Hankuk Aviation Univ.

RSP Lab

PRF Staggering
- Consider a radar system with two interpulse periods T1 and T2
T1 n1

(9.57)
T2 n2
Where, n1 and

n2 are integer

- The first true blind speed occurs when


- The ratio (stagger ratio)

Hankuk Aviation Univ.

n1 n2

T1 T2

ks

n1
n2

(9.58)
(9.59)

RSP Lab

PRF Staggering

Using staggering ratios closer to


unity the first true blind speed
farther out

- The dip in the vicinity of 1 / T1


becomes deeper
< Frequency responses of a single canceler. T1=4, T2=3, T1/T2=4/3 >
Hankuk Aviation Univ.

RSP Lab

PRF Staggering

< MTI responses, staggering ratio 63/64 >

- In general, if there are N PRFs related by


n
n1 n2

N
T1 T2
TN

(9.60)

- If the first blind speed to occur for any of the indiviual PRFs is vblind1
- The first true blind speed for the staggered waveform is
vblind
Hankuk Aviation Univ.

n1 n2 nN
vblind1
N

(9.61)
RSP Lab

9.11 MTI Improvement Factor


Performance of MTI systems
- Clutter Attenuation (CA)
- MTI Improvement factor
(1) MTI Clutter attenuation
CA = Ci / Co
Ci : MTI filter input clutter power
Co : Output clutter power

(9.62)

(2) MTI Improvement factor

S0
I
C0

Si S 0 Ci

Ci Si C0

S0
CA
Si

Hankuk Aviation Univ.

(9.63)

(9.64)
So/Si = |H(w)|2 : average power gain for MTI filter
RSP Lab

MTI Improvement Factor


Gaussian clutter power spectrum

W( f )

Pc
exp( f 2 2 c2 )
2 c

Pc : clutter power (constant)

(9.65)

c : clutter rms frequency

c 2 v /

(9.66)

v : rms wind velocity => wind: main reason of clutter freq. spreading

f 22
Pc

W( f )
exp
2
8
2 2 v

(9.67)

Clutter power at the input of an MTI filter

Co

Hankuk Aviation Univ.

f2
exp 2 df
2 c
2 c
Pc

(9.68)

RSP Lab

MTI Improvement Factor


Factoring out

Ci Pc

f2
exp 2 df
2 c
2 c
1

(9.69)
(9.70)

Ci Pc
Clutter power at output of an MTI

Co W ( f ) H ( f ) df
2

(9.71)

Analysis using a single delay line canceller


Single canceller power gain

f
H ( f ) 4 sin
fr
2

Hankuk Aviation Univ.

(9.72)

RSP Lab

MTI Improvement Factor


Eq.(9.71) & Eq.(9.72)

Co

f 2 f
Pc
exp 2 4 sin
2 c
2 c f r

df

(9.73)

Small f, then ratio f/ fr is very small. (ie c<<fr)


Using the small angle approximation

Co

f f
exp 2 4
2 c
2 c f r
Pc

Rewritten as

4 Pc 2
Co
f r2

df

(9.74)

f2 2
exp 2 f df
2 c
2 c
1

(9.75)

Integral part is the second moment of a zero mean Gaussian distribution


2
2
4
P

with c
C c 2
o

Hankuk Aviation Univ.

2
r

(9.76)

RSP Lab

MTI Improvement Factor


Substituting eq.(9.76)&(9.70) into (9.62)
2

CA

Ci f r


Co 2 c

(9.77)

Improvement factor for a single canceller

S0 f r

I
Si 2 c

(9.78)

Power gain ratio for a single canceller


2

f
sin df
4

fr
fr / 2
2
Using the trigonometric identity (2 2 cos 2 ) 4(sin )
f /2
1 r
2f
2

df 2
H( f )
2 2 cos

fr fr / 2
fr
So
1
2
H( f )
Si
fr

Hankuk Aviation Univ.

fr / 2

(9.79)

(9.80)

RSP Lab

MTI Improvement Factor


It follows that

fr

I 2
2 c

=> Only c<<fr

(9.81)

When c<<fr is not true, the autocorrelation function needs


Ex 9.3
fr =800Hz, clutter rms c =6.4Hz (Wooded hills v = 1.16311 km/hr)
find the improvement factor when a single delay line canceller is used.
Solution : clutter attenuation (CA)
2

f 800
395.771 25.974dB
CA r
2 c 2 (6.4)
since S0/Si = 2 = 3dB
IdB = (CA + So/Si) dB = 3 + 25.97 = 28.97dB
Hankuk Aviation Univ.

RSP Lab

9.12 Subclutter Visibility (SCV)


Phrase Subclutter Visibility (SCV)
- radars ability to detect non-stationary targets in a strong clutter background
- used as a measure of MTI performance

<Illustration of SCV (a) MTI input (b) MTI output>


Hankuk Aviation Univ.

RSP Lab

Subclutter Visibility (SCV)


Phrase Interclutter Visibility (ICV)
radars ability to detect non-stationary targets between strong clutter points
- > if radar system - resolve the area of strong and weak clutter
SCV - expressed as the ratio of the improvement factor to the min.MTI output
SCR
(9.82)
SCV I /( SCR)o
SCV of two radars not compare their performance
-> target-to-clutter ratio : proportional to the size of the radar resolution
cell and may also be a function of frequency.
ex) Radar system with 10us pulse length & 10o beamwidth : need 20dB more
SCV than Radar system with 1us pulse length & 1o beamwidth.
Hankuk Aviation Univ.

RSP Lab

9.13 Delay Line Cancellers with Optimal Weights


Delay line canceller transversal FIR filter (tapped delay line filter)

<N-stage tapped delay line implementation>

Weights : binomial coefficients -> N-stage cascaded single line cancellers


Binomial coefficient :

wi (1)

i 1

Hankuk Aviation Univ.

N!
; i 1,..., N 1
( N i 1)!(i 1)!

(9.83)

RSP Lab

Delay Line Cancellers with Optimal Weights


Using the binomial coefficients produces an MTI filter (approximated optimal
filter)
-> maximize the improvement factor

Two equivalent three delay line cancellers


(a) Tapped delay line (b)Three cascaded single line cancellers
Hankuk Aviation Univ.

RSP Lab

Delay Line Cancellers with Optimal Weights


Average power gain for an N-stage delay line canceller

f
S0
2
H1 ( f ) 4 sin
Si i 1
i 1
fr
For example, N=2 (delay line canceller)
N

Rewritten

f
S0
16 sin
Si
fr
S0
2N
H1 ( f )
Si

(9.84)

f
2N
2 sin
fr

(9.85)

2N

(9.86)

Blind speeds for N-stage delay canceller : identical to single cancellers blind speed
Blind speed : independent from the number of cancellers used

S0
N ( N 1) N ( N 1)( N 2)
1 N 2

...
Si
2!
3!

Hankuk Aviation Univ.

(9.87)

RSP Lab

Delay Line Cancellers with Optimal Weights


General expression by Nathanson

( S o / Si )
I N N
* (k j )

wk w j

k 1 j 1
fr
wk & wj
((k-j)/fr)

(9.88)

: weights of tapped delay line canceller


: correlation coefficient between the kth and jth samples

For example, N = 2

Hankuk Aviation Univ.

1
4
1
1 T 2T
3
3

(9.89)

RSP Lab

You might also like