Pri Analysis and Deinterleaving by DR Wiley PDF
Pri Analysis and Deinterleaving by DR Wiley PDF
Pri Analysis and Deinterleaving by DR Wiley PDF
2
PULSE REPETITION INTEVAL (PRI)
The effects of the one-way range equation of ELINT and the two-
way range equation of radar on signal strength must be understood
and explored in order to appreciate the typical situations
encountered in ELINT and EW. Similarly, the constraints placed on
radar waveforms must be understood in order to correctly interpret
the functions and applications of the signals transmitted by radar
and also to be aware of the signal characteristics expected to be
encountered by ELINT. In many ways, understanding these aspects
of ELINT is what separates one who only observes signals from one
who both observes and analyzes signals.
4
Radar and ELINT Range Equations
PT GT G R O V 2
SR
(4S ) R R LT LR
3 4
PT GTE G E O2
SE
(4S ) R E LT LE
2 2
100
in b
Ma
RangeRatioSLi
dB
RE/RR
RangeRatioMBi E
=0
: GT A
e
e lob
10 Sid
V 1 sq. m
G 30 dB
R
G 100
G 1
E
1 3
1 10 100 1 10
Ri
Range (km)
Figure 2-1 ELINT to Radar Range Ratio
3
1 10
Resolution (meters)
(meters)
100
RangeResolution
RngRes bi
Range
10
1
6 7 8
1 10 1 10 1 10
bi
Bandwidth B(MHz)
Bandwidth (MHz)
T ACC
v v 2a(GR)
2
0.5
v v 2a('R)
2
0.5
a 12
a
Constraints on Time-Bandwidth Product or Pulse Compression
Ratio
Because range resolution is determined by bandwidth and integration
time is determined by velocity, there is a natural limit on the product
of the instantaneous bandwidth and the duration of the coherent
processing interval or pulse width. This is called the "time-
bandwidth product." The radar's pulse compression ratio is limited to
no more than its time bandwidth product. By combining Equations
for range resolution and integration time it is easy to see that the time
bandwidth product is limited to:
Bv § ac · c
BT ¨¨ 1 2 1¸¸
a o0
o
a © Bv ¹ 2v
13
6
1 10
BT Limit product BT
BT i 1 a=0 g
BT i 2 a=1 g
Maximum time-bandwidth
g
BT i 5 a=2
g
BT i 10 a=5
g
BT1 i 10
a=
5
1 10
4 5 6
1 10 1 10 1 10
bi
Signal Bandwidth B (Hz)
Bandwidth
Figure 2-414
Limit on Time x Bandwidth
Constraints on Doppler Resolution
If the radar coherently integrates the echoes in one range cell for the
entire integration time, the minimum doppler filter bandwidth, Bf, is
approximately the reciprocal of the integration time,.T, which is
either TCV for constant velocity targets or TACC for accelerating
targets:.
1
Bf |
T
However if the target is accelerating, the doppler shift changes.
Clearly there is a relationship between acceleration and the time the
doppler shift of the moving target remains within the doppler filter
bandwidth.
2aTf o 2aT
'f acc Bf
c O
15
O 2a
T , Bf !
2a O
16
4
1 10
3
6.502 u10
3
a=10g
1 10
Doppler Spread( kHz)
'f i 1
100
'f i 2
a=1g
'f i 5
10
'f i 10
0.65
0.1 3
1 10 0.01 0.1 1
3 Ti 1
1u10
17
4 3
1 10 1 10
3 1000
6.502u10
a=10g
le
t sca
3
1 10
a=5g
ad - ri gh
r Sp r e a=2g g
ple a=1
Doppler Spread( kHz)
100
Dop
'f i 1
Bandwidth (MHz)
100
'f i 2
Bi
'f i 5 Ma
10
xi m
'f i 10 um
Sig 10
n al B
and
w idt
1
h -le
f t sc
ale
0.65 1
0.1 3 1
1 10 0.01 0.1 1
.001 Ti 1
18
The doppler filter bandwidth must be no wider than the spread of
doppler frequencies expected. Figure 2-5 also shows the
maximum radar signal bandwidth. For the case where acceleration
has a minimal effect on the integration time, the maximum
acceleration of the target can be expressed in terms of the radar
signal's bandwidth as
2
v
a max 2B 2
c( RF )
19
20
Frequency Agility
From one coherent processing interval to the next, the radar can
change its carrier frequency without changing its range resolution
properties. The agility band is limited by the radar designer’s
ability to obtain sufficient power and to maintain beam width and
pointing angle--typically about 10% of the center frequency. (For
example, a 1 GHz agility band centered at 10 GHz.) What this
means for ELINT is that narrowband receivers have a low
probability of intercepting the complete radar transmission. If it is
sufficient to intercept only portions of the radar transmission,
narrowband receivers can be slowly tuned across the radar band
and the entire agility band can still be determined if the signals is
present for enough time. The coherent processing interval
determines the Doppler resolution. When FA is used with doppler
processing, the frequency is changed on a pulse-burst to pulse-
burst basis, not a pulse-to-pulse basis.
21
PRI Agility
Modern multifunction radar systems make use of multiple pulse
repetition intervals (PRI) values during one look at the target. It is
a requirement of today’s pulse doppler radars that the PRI remain
constant during each coherent processing interval. For moving
target indicating (MTI) radar designs, there is usually a sequence
of PRI values that must be completed during one processing
interval. This repeated sequence is known as "stagger" and ELINT
analysts call the period of the stagger the stable sum. This is
because when consecutive PRIs are added, the sum is constant
when one adds together the PRIs which make up the stagger
period--regardless of which PRI is selected as the starting point for
the sum.
22
MTI radars operate by subtracting (in amplitude and phase) the
echoes from one PRI from those in the next PRI. Stationary targets
have the same phase and amplitude and thus “cancel.” Echoes
from moving targets generally do not have then same amplitude
and phase and so do not cancel. However if the target moves an
integer multiple of half wavelengths in one PRI, the phase of the
second echo is shifted by a multiple of 360 degrees from the first
and the echoes cancel. Such speeds are “blind speeds.” Changing
the PRI changes the blind speed. A PRI sequence is selected to
detect targets regardless speed Moving target detection (MTD)
radar systems use a doppler filter bank to divide the frequency
region between the PRF lines into several filter bands (for example:
8 bands). This requires repeated constant PRIs (say 10 pulses at
one PRI and then 10 pulses at another, etc.) Multiple PRIs are
required due to range and velocity ambiguities and make visible
target ranges and velocities “eclipsed” by transmitted pulses (in
time) or spectral lines (in frequency).
23
For constant PRI and RF, the maximum unambiguous range (Ru)
and the maximum unambiguous velocity (Vu) are given by:
c(PRI ) c
Ru Vu
2 2( RF )( PRI )
Examples at 10 GHz:
PRI 1000 us, Vu=15 m/s and Ru=150 km
PRI 100 us, Vu=150 m/s and Ru=15 km
PRI 10 us, Vu=1500 m/s and Ru=1.5 km
1000
3
1 10
3 4
10 100 1 10 1 10
Unambiguous
25 Velocity (m/s)
Fi 2 7R /V l it R l t d
Frequency
Agility Band
(Depends
on
Frequency
Component
* Design,
ECM
Factors,
Designer
Ingenuity)
Time
*BandwidthDepends
Determines Range Resolution Which
on Radar Mission
Figure 2-8. Modern frequency Agile Radar with 100% Duty Factor
26
USES OF PRI
27
Analysis p. 147
28
RANGE-VELOCITY AMBIGUITY
Analysis p. 148
29
30 Text p. 149
OPTIMUM PRI FOR MEDIUM PRF RADAR
Band “Be” Obscured at each PRF line
31
32
PRI DRIFT
33 Analysis p. 153
35
36
PRI STAGGER
• Desired Parameters
- Number of intervals
T T T T - Number of positions
Unmodulated Pulse Train - Interval values
- Sequence
H H
- Stable sum
37
12. 6.7 4000 3571.4 (3) Exact order of 1 pulse 3 pulse canceller for MTI. Stagger to Surveillance
4405.1 (3) intervals is not known eliminate blind speeds
3745.3 (3)
4255.3 (3)
4081.6 (3)
13. 1-100 400 For first sequence only: Various Sequences To eliminate blind speeds. Has various Surveillance, tracking, kill
62.1 623.3 16:21:19:17:20:18 digital MTI processing including double assessment, missile guidance
2500 818.0 16:17:16:17 double-cancellation
740.1 16:19:16:19 38
662.0 16:21:16:21
701.1 16:17
DESCRIPTION OF PRI VARIATIONS
Type 1 2 3 4 5 6 7 8
39
JITTERED PRI
• Intentional Jitter
- Discrete or continuous
• Desired Measurements
- Mean PRI
- Peak PRI deviation limits
- PRI distribution (histogram)
- Number of discrete PRIs
40
RADARS WITH JITTER
Pulse PRI Peak-to-Peak Jitter Peak-to-Peak Jitter Jitter Type Jitter Purpose Radar Function
Width (μs) (μs) (%)
(μs)
Definition: Rapid (automatic) switching between discrete PRIs with a dwell at each PRI
PRI = T1 PRI = T2
• Desired measurements
- Number of PRIs
- Value of PRIs
- Dwell times
- Total dwell time for sequence
- Dwell sequence
- Time to switch
42
SLIDING PRI
Definition: The pulse train has a PRI (PGRI) that is continuously changing in either
a monotonically increasing or decreasing manner between maximum
and minimum PRI limits.
• Desired Parameters
43
• Periodic Modulation
- Displacement value
44
OTHER PRI TYPES 2
- Range of on-period
- range of off-period
Definition: Pulse train that is transmitted for some purpose for a relatively short
time and then is off for a relatively long time
- Burst definition
- Number of bursts per second
- Relationships of burst to scan
45
SCHEDULED PRIs
• Scheduled PRIs
Definition: PRIs are computer controlled, vary with the target environment and
function being performed by radar, and cannot be described by other
definitions
- Number of intervals
- Interval values
- Typical sequences
- Reason for sequence
46
MUTLIPLE PULSE GROUPS
• Constant and Cyclic Patterns
47
Analysis p. 151
48
DOPPLER EFFECT
v = radial velocity
c = 3(108) m/sec
fo = transmitted RF
2v
Doppler Shift f f fo r x fo
d 1 c
49
50
FOURIER TRANSFORMS
51
52
FM THEORY
V(t) A sin(2ʌ f c t ș(t))
Phase Disturbance
Total Phase 2ʌ f c t ș(t)
1 d 1 dș
Instantaneous Freq. (total phase) fc
2 dt 2 dt
1 dș
f m ǻijcos2ʌf m t
2ʌ dt
1 dș
ǻFcos2ʌf m t
2ʌ dt
THEN :
ǻf
V(t) Asin(2ʌ f c t sin2ʌ f m t)
fm
INDEX OF
MODULATION
53 " M"
BESSEL EXPANSION
J (m) . . . . .
3
J0(m)
J1(m) J1(m)
J2(m) J2(m)
fc-2fm fc fc+2fm
fc-fm fc+fm
54
BESSEL FUNCTIONS
55
ǻf
m 1 i.e. ǻf is small
fm
m
J o (m) | 1 J (m) |
1 2
THEREFORE
m m
V(t) # A[sinȦc t sin(Ȧc Ȧm )t sin(Ȧc Ȧm )t]
2 2
A
V m ǻf
SB
Vc 2 2f m
mA/2 mA/2
in dB
V ǻf
fc-fm fc fc+fm 20 log SB 20 log
Vc 2f m
56
EXAMPLES
20 log ǻf 40 dB
2f m
ǻf 1 kHz
2f m
e.g. f m 1 kHz
ǻf 20 Hz
IF f c 10 GHz, STABILITY IS
20 Hz 2 parts in 109
10 x 109 Hz
57
12 μs = X
T1 = 40 μs
2 μs = Y
T2 = 30 μs
N1 T1 + X = T and N2 T2 + Y = T
59
60
ERICSSON PS-05/A MULTI-MODE OPERATION (2)
61
MTI VIDEO
62
MTI PHASE SHIFTS
63
64
BIPOLAR VIDEO
65
DOPPLER RETURNS
TRAIN
CAR
Typical images displayed on TPS-25 ground
Surveillance radar. Shown are target images
of: 1) a train, 2) an automobile, 3) a walking man,
and 4) a walking girl. (US Army photograph.)
MAN WALKING
WOMAN
WALKING
66
PULSED-OSCILLATOR MTI
Zeros at 0, ʌ, nʌ when
n
f
d T
cxn
so blind speeds are V
b 2RF x T
nȜ
x PRF
2
Barton, p. 192
67
Page M50.ppt
68
BLIND SPEED ELIMINATION
No Stagger vb = n c/2(PRI)(RF)
1
T
Tİ 63
Tİ 65
Null at
64/T
1 1 1 1 ....
I I I I
1 2 3
I1, I2, I3 are the individual improvement factors calculated on basis of PRI, pulse
amplitude, pulsewidth, transmitter frequency, ……….. stabilities
70
INSTABILITY LIMITATIONS
71
CLUTTER STRENGTH
72
MTI + PULSE DOPPLER = MTD
8-Pulse Weighting
3-Pulse
Doppler And
Canceller
Filter Bank Magnitude
73
MTD PERFORMANCE
• Theoretical
• Practical
Performance of FAA ASR radar: 3 pulse MTI alone | 25 dB
3 pulse MTI + 8 pulse FFT | 45 dB
75
PRI EXERCISES
1. The analyst found a signal at 6 GHz which had two-interval, two-position stagger. The
intervals were 500 and 550 microseconds. What is the average PRI? What is the
stagger ratio? What is H? What are the new blind speeds?
2. What is the improvement factor for MTI of a radar which has RMS jitter of 10 nanosec
and a pulse duration of 1.41 microsec?
3. A discrete random jitter PRI train was analyzed and the PRIs were found to be one of
the following 5 nominal values:
Nom
PRI (μsec)
2440.8
2428.7
2465.3
2453.1
2562.9
Is there a clock? If so, what countdowns are used and what is the clock frequency or
period? What common range mark is that closest to?
76
PRI EXERCISES #2 - ANSWERS
1. (500 + 550)/2 = 525 microsec = average PRI
R = 550/500 = 1.1 (11:10)
H = 550 – 525 = 25 microsec
Blind speed before stagger = nc/(2 • PRIave • RF)
(3 x 108 ) m / sec
VB 171.4 km / hr (106.5 mph )
2(525)(10 6 ) sec x 6(109 ) x 1 / sec
77
T
Triggering Error 'T RISE x 'A
0.8 A
'A A
'T
78
PRI VARIATION DUE TO NOISE
V2
amplitude Noise Power 1
(Amplitude)2 Signal Power SNR
T
V Rise 1
Time 0.8 SNR
V2 V2 V2 2 V2
PRI Time1 Time 2 Time
T
V 2 Rise
PRI 0.8 SNR
79
2
§
t · .35
SNR 3.125 ¨¨ V r ¸¸ tr |
¨
© PRI ¹
¸ Bandwidth
0.1 3.5 Ps 81 dB 61 dB 41 dB
1.0 0.35 Ps 61 dB 41 dB 21 dB
10.0 35 ns 41 dB 21 dB X
100.0 3.5 ns 21 dB X X
80
AMPLITUDE INDUCED ERROR
81
82
PERFORMANCE OF TRIGGER CIRCUIT
83
84
DELAY AND PULSE JITTER
Delay D2
Delay D1
85
~ 1 μs Jitter ~ 2 μs Jitter
Delay = 1 PRI Delay = 5 PRI
86
SYNTHESIS OF AVERAGE PRI
87
88
REAL TIME RASTER DISPLAYS
89 Analysis p.74
90
DTE MODE-CIRCULAR SCAN
RADAR
91
RTR SIMULATION ON A
PERSONAL COMPUTER
92
MEAN PRI ESTIMATES
93
94
RMS ERRORS COMPARED
95
96
97
NONCUMULATIVE AND
CUMULATIVE JITTER
98
CRAMER-RAO BOUNDS COMPARED
99
100
USING THE WRONG JITTER
MODEL
101
PRI HISTOGRAMS
102
ACTIVITY IN 0.1S INTERVALS
103
INTERVALS FORMED BY
PULSE PAIRS
104
DELTA-T HISTORGRAM
(10% JITTER)
105
DELTA-T HISTOGRAM--
STAGGER
••• 4 5 7 4 5 7 4 5 •••
t0 = 0 t1 = 4 t2 = 9 t3 = 16 t4 = 20 t5 = 25 t6 = 27 t7 = 31 t7 = 37
A. (tn – tn-1) = 4, 5 or 7
B. (tn – tn-2) = 9, 11 or 12 (4 + 5, 4 + 7, 5 + 7)
C. (tn – tn-3) = 16 (4 + 5 + 7)
F. (tn – tn-6) = 32
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32
A B C D E F
106
THREE POSITION STAGGER
107
DELTA-T HISTOGRAM:
TOA AUTOCORRELATION
n
f (t ) ¦ G (t t n ) .....
n 1
t1 t2 . . . . . t3 t4 . . . . .
f
h (W) ³ f (t ) f t W dt
f
f ª º
h (W) ³ «¦ G (t t n ) ¦ G (t t k W)»» dt
f «¬ n k ¼
t=0 W
108
t = t1 + W
DELTA-T HISTOGRAM:
TOA AUTOCORRELATION
h(W) ¦ ¦ G(t n t W)
n k k
IJ IJ
2 2
³ h(W) ³ ¦ ¦ G(t n t k W)
IJ IJ n k
1 1
THEREFORE:
A count of the number of pairs of pulses whose arrival
times differ by a value between W1 and W2 is equal to
the integral of the autocorrelation of the TOA’s
109
Center Frequency
(average PRF)
Jitter
Waveform FM Trigger Time of
Oscillator Generator Arrival
Peak Sequence
Amplitude
Periodicities
• Periods
• Amplitudes
¦ Drifts/Trends
• Slopes
Random Components
• Bandwidths
• Variances
110
• Probability Densities
INSTANTANEOUS
FREQUENCY ESTIMATION
2500 Linear
2000 2000 2000 Interpolation
Freq
1428.5 1666.7
Midpoints
of Intervals
111
DEINTERLEAVING DEVICE
112
DEINTERLEAVING VIA DELTA-W HISTOGRAM
113
114
NUMBER OF EMITTERS DEINTERLEAVED
115
116
COMPLEX DELTA-W HISTOGRAM - II
117
118
EFFECT OF A NEAR MULTIPLE PRI
119
120
Delta-T Histogram for Ten Interleaved Pulse Trains
Delta-T Histogram
100
Histogram Count
dhist b
50
.75 max( dhist )
0
5 4 4 4 4
8 10 1 10 1.2 10 1.4 10 1.6 10
6
int vb PRI k 10
PRI, Seconds
121
100
50
100
0
4 4 4 4
1 10 2 10 3 10 4 10
int vb PRI k 10
6
int v PRI 10
b k
6
abchist b dhist b
50 0
1.05 max( abchist) 1.05 max( dhist )
50
0 100
5 4 4 4 4 4 4
5 10 1 10 1.5 10 2 10 2.5 10 3 10 3.5 10
intv b PRI k 10
6
intv PRI 10
b k
6
PRI, Seconds
7
Jitnc 0.5 Jitcum 0.5 N 820 width 5 u 10 10 Interleaved Pulse Trains
0 0
123
50
abchist b dhist b
50 0
1.05 max( abchist) 1.05 max( dhist )
50
0 100
5 4 4 4 4 4 4
5 10 1 10 1.5 10 2 10 2.5 10 3 10 3.5 10
intv b PRI k 10 intv PRI 10
6
b k
6
PRI, Seconds
7
Jitnc 1 Jitcum 1 N 820 width 5 u 10 10 Interleaved Pulse Trains
0 0
124
Effect of Jitter on Delta-T Histograms (Jitter=5 microseconds)
abchist b dhist b
50 0
1.05 max( abchist) 1.05 max( dhist )
50
0 100
5 4 4 4 4 4 4
5 10 1 10 1.5 10 2 10 2.5 10 3 10 3.5 10
intv b PRI k 10
6
intv PRI 10
b k
6
PRI, Seconds
7
Jitnc 2.5 Jitcum 2.5 N 820 width 5 u 10 10 Interleaved Pulse Trains
0 0
125
Uniform Jitter=0.002
Uniform Jitter=0.02
0 0
0
0 1· 10-4 0 1· 10-4 0 1· 10-4
1 1. 048· 10-4 1 1. 05·10-4 1 1. 05·10-4
2 1. 11·10-4 2 1. 11·10-4
2 1. 11·10-4
3 1. 15·10-4 3 1. 15·10-4 3 1. 15·10-4
4 1. 162· 10-4 4 1. 164· 10-4 6
PRI 10 4 1. 163· 10-4
5 1. 176· 10-4 5 1. 178· 10-4 5 1. 177· 10-4
6 1. 19·10-4 6 1. 192· 10-4 6 1. 191· 10-4
pk 7 1. 21·10-4 pkc 7 1. 21·10-4 7 1. 21·10-4
8 1. 23·10-4 8 1. 23·10-4 8 1. 23·10-4
9 1. 26·10-4 9 1. 26·10-4 9 1. 26·10-4
10 0 10 0
11 0 11 0
12 0 12 0 This example based on the method of B.
Frankpitt, J. Baras, A. Tse, "A New Approach
13 0 13 0 to Deinterleaving for Radar Intercept
14 0 14 0 Receivers," Proceedings of the SPIE, Vol
5077, 2003, pages 175-186
15 0 15 0
PRF Spectrum
0.01
Amplitude
Xj
0.00011 max( X)
0.005
0
4 4 4 4 4 4
6000 8000 1 10 1.2 10 1.4 10 1.6 10 1.8 10 2 10
PRF Spectrum
0.03
Xj
Amplitude
.025 0.02
.015
0.01
0
4 4 4 4 4 4
4000 6000 8000 1 10 1.2 10 1.4 10 1.6 10 1.8 10 2 10
f j 1 PRF k 2 PRF k
PRF (Hz)
3
10 Interleaved pulse Trains N 1.741 u 10
129
PRF Spectrum
0.03
Xj
Amplitude
.03 0.02
.02
0.01
0
4 4 4 4 4 4
4000 6000 8000 1 10 1.2 10 1.4 10 1.6 10 1.8 10 2 10
f j 1 PRF k 2 PRF k
PRF (Hz)
10 Interleaved pulse Trains N 871
131
100
90 Amp Off n
pO
Am
80
1% Jitter
ff
70 pO
A m
60
Score n
50 pO
Simulated Data Am
Average Density 200 pps
40
Amp Off
30
20
8% Jitter
10
135
^ĐĞŶĂƌŝŽ
ŽŵƉůĞdžĞůƚĂͲd,ŝƐƚŽŐƌĂŵWĞĂŬƐ /ŶƉƵƚWZ/Ɛ
136
WƵůƐĞƚĂŐƐĨŽƵŶĚŝŶƉĞĂŬ
ηϮ
137
WƵůƐĞdĂŐƐŝŶWĞĂŬηϮǁŚŝĐŚŚĂǀĞƚŚĞ
ŵŽƐƚƉƌĞǀĂůĞŶƚWZ/WŚĂƐĞ
138
PRI ANALYSIS EXERCISE
Two signals are observed with the same angle of arrival but on different frequencies. The PRI of one is nearly
stable at 3000 μs. The PRI of the second jitters randomly with a mean value of 1500 μs and a peak-to-peak jitter
of about 20 μs. The analyst notices that the PRI’s of the second signal can be paired such that their sum is nearly
stable at 3000 μs; i.e., PRI #1 + PRI #2 = PRI #3 + PRI #4 = PRI #5 + PRI #6, etc. However, PRI #2 + PRI #3 z
PRI #4 + PRI #5. He also notices that the mean value of the second signal’s PRI is exactly one-half that of the
first signal’s PRI every time the two signals are reported. The first signal has a slow circular scan, the second a
faster sector scan. What conclusions might be drawn about these two radars?
What additional data would you request from the ELINT station?
139
There is a good possibility that the second radar operates in PRI synchronism with
the first;
but at one-half the PRI. Alternate pulses are triggered by the master clock, the
intermediate
pulses are generated by “one shot” type delay circuit which is not stable.
The second radar may be a height finder using elevation sector scan and associated
with a long
range search radar.
Confirmation of this would be aided by using two receivers and making a recording
of both
Signals simultaneously to investigate whether the second signal is synchronized to the
first.
140
PRECISION PDWs
'pw 't 2 2
r 'tf
where 'pw RMS error in pulse width
't r RMS error of pulse risin g edge time
't RMS error of pulse falling edge time
f