Sie sind auf Seite 1von 7

Journal of the Chinese Institute of Engineers

ISSN: 0253-3839 (Print) 2158-7299 (Online) Journal homepage: https://www.tandfonline.com/loi/tcie20

Deception jammer rejection using pulse diversity


in joint slow/fast-time domain

Gang Lu & Bin Tang

To cite this article: Gang Lu & Bin Tang (2013) Deception jammer rejection using pulse diversity
in joint slow/fast-time domain, Journal of the Chinese Institute of Engineers, 36:3, 405-410, DOI:
10.1080/02533839.2012.731847

To link to this article: https://doi.org/10.1080/02533839.2012.731847

Published online: 15 Oct 2012.

Submit your article to this journal

Article views: 151

Citing articles: 6 View citing articles

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=tcie20
Journal of the Chinese Institute of Engineers, 2013
Vol. 36, No. 3, 405–410, http://dx.doi.org/10.1080/02533839.2012.731847

Deception jammer rejection using pulse diversity in joint slow/fast-time domain


Gang Luab* and Bin Tangb
a
School of Electrical and Information Engineering, Xihua University, Chengdu 610039, China; bSchool of Electronic
Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
(Received 29 April 2011; final version received 22 November 2011)

In this article, a method based on the theory of pulse diversity is presented for rejecting a deception jammer which
is repeated by a digital radio frequency memory. The application conditions involved in slow-time approach,
which cancels the jammer across several pulse repetition intervals, is restricted to the hypothesis that the reflectors
and reflectivity levels remain stationary making it impractical in rapidly varying conditions. This issue is resolved
through producing new vectors from the original transmission signals in joint slow/fast-time domain. Moreover,
the presented method also insures satisfactory performance levels when the target and jamming pulses are
overlapped.
Keywords: electronic counter-countermeasures; digital radio frequency memory; false target; space time
block code

1. Introduction The results are that the associated conditions imposed


In electronic warfare, the false target is a commonly are lifted and the method is practicable in rapidly
used electronic countermeasure (ECM) method to varying conditions since no cancellation is required
deceive surveillance radar systems (Pace et al. 2002, across different PRIs. Furthermore, the non-over-
Berger 2003). These kinds of jamming, generally lapping requirement is also eliminated.
produced by a modern ECM facility equipped with a
digital radio frequency memory (DRFM), pose a
difficult and challenging environment for victim 2. Model and algorithm formulation
radars. A processing technique that may be referred
2.1. Whitening-matched filter
to as pulse diversity can overcome this problem
(Soumekh 2006, Akhtar 2009, Lu et al. 2010). In We assume here that the complex vector measured by
such processing, one needs several pulse repetition the radar during the time on target is given by
intervals (PRIs) in order to form a full block code or  T
z ¼ zð0Þ zð1Þ    zðKÞ , ð1Þ
phase perturbations in slow-time and the jammer can
be rejected or mitigated by a series of sophisticated where K is the number of samples during a single pulse.
processes. The techniques can successfully reject the It is assumed that the recorded samples consist of the
false targets provided that the reflectors and reflectivity sum of target ps and disturbance d, i.e.,
levels remain stationary across a full code period
z ¼ ps þ d, ð2Þ
(Akhtar 2009, Lu et al. 2010) or the probability of real
reflectors overlapped with jamming pulses is small where s is the signal vector and p the complex
(Soumekh 2006, Akhtar 2009). amplitude. The disturbance d is defined as the sum of
In this study, efforts are made to remove the white thermal noise n  CNð0, n2 IÞ and correlated
stationary hypothesis in the presence of false targets. Gaussian clutter c  CNð0, c2 RÞ, where I denotes the
For the purpose of this, a new transmission pulse block identity matrix, R the normalized clutter covariance
is generated from the original radar signals. In contrast matrix, and n2 and c2 the variances of n and c,
to the slow-time method, a dual sub-code pulse in fast- respectively. Consequently, d is modeled as a complex
time domain is transmitted from radar to implement a Gaussian distributed random vector with zero
space time block code (STBC) in half quantity of PRI. mean and covariance matrix d2 M ¼ EðzzH Þ, where

*Corresponding author. Email: aeneas999@hotmail.com

ß 2013 The Chinese Institute of Engineers


2406 G. Lu and B. Tang

d2 ¼ n2 þ c2 is the total disturbance power and M where si(t) (i ¼ 1, 2) denotes the radar transmission
the normalized disturbance covariance matrix formu- signal vector. Actually, we can deem si(t) as a dual
lated as vector, i.e.,
 T
n2 I þ c2 R 1 CNR si ðtÞ ¼ sTi1 ðtÞ sTi2 ðtÞ : ð8Þ
M¼ ¼ Iþ R: ð3Þ
n2 þ c2 CNR þ 1 CNR þ 1
When radar signals are received by a jammer, the
Denote by CNR the clutter-to-noise power ratio, more jammer needs to detect and analyze the incoming
precisely, CNR ¼ c2 =n2 : In shorthand notation, pulses in order to handle the valid signals and produce
d  CNð0, d2 MÞ: Besides, we suppose that a set of effective jamming pulses. It requires a certain amount
secondary data free of signal components but sharing of time to accomplish the whole process with updated
the same statistical properties of the disturbance in the pulses. It is then possible for the radar to cope with the
cell under test is available. repeated jamming signal recorded in a DRFM in
Actually, M is Hermitian and we also assume that previous PRI (Soumekh 2006, Akhtar 2009, Lu et al.
M is full rank (this condition is usually satisfied in 2010). Moreover, even if a jammer could detect and
practice). Therefore, it can be factored as M ¼ LLH , re-transmit the current pulse at once, it would have to
where L is a lower triangular matrix. The output of the repeat the pulse recorded in previous PRI when it
whitening-matched filter (WMF) is given by (Gini and attempts to protect targets which are closer to the
Farina 2002) victim radar than itself. As a consequence, it is
assumed that the jammer lags one PRI behind the
FðzÞ ¼ sH M1 z ¼ zH M1 s
victim radar. If the delayed PRI is greater than one, a
¼ ðL1 zÞH ðL1 sÞ: ð4Þ higher order STBC (Tarokh et al. 1999a, 1999b) can be
utilized to overcome this challenge. Suppose the
In Equation (4), L1 z denotes the new whitening
jamming pulses are given by
received vector, while L1 s the new whitening trans-
mitting signal replica of radars. If the filter realization    T
jðtÞ ¼ T sT0 ðtÞ T sT1 ðtÞ , ð9Þ
form is used in a radar, the corresponding impulse
response of the WMF is h ¼ s. Then at time t, the where T is a linear operator which denotes the process
WMF output is given by hH M1 zðtÞ where M1 zðtÞ or transformation performed on the radar signals to
denotes the transformed input signal vector at time t. produce the false target. In the initial pulse diversity
block, s0(t) may be any arbitrary pulse (signal stored in
a DRFM of jammers). In subsequent blocks, however,
2.2. STBC pulse sequence it is determined (just the last transmission pulse of the
victim radar). The received pulses at the radar side can,
We formulate the jammer rejection method as a
therefore, be modeled as
temporal pulse diversity problem on the STBC Q
8 P
(Xu and Kwak 2005) >
> z1 ðtÞ ¼ m ð1Þs1 ½t  m ð1Þ
>
> m
X
M  >
> P 0   
 >
> þ n ð1ÞT s0 t  n0 ð1Þ þ d1 ðtÞ,
Q¼ qm Am þ qm Bm , ð5Þ <
n
m¼1 P ð10Þ
>
> z 2 ð tÞ ¼ m ð2Þs2 ½t  m ð2Þ
>
>
where Am and Bm are the N  Nt matrices and fqm gM m¼1
>
>
>
m
P   
the transmission signal set. In particular, Q is exactly >
: þ 0n ð2ÞT s1 t  n0 ð2Þ þ d2 ðtÞ,
n
the Alamouti (1998) code if N ¼ Nt ¼ M ¼ 2 and
8     where m(i) and 0n ðiÞ (i ¼ 1, 2) denote the complex
>
> 1 0 0 1
> A1 ¼
< , A2 ¼ , amplitude of the target and jamming signal, respec-
0 0 0 0 tively,  m(i) and n0 ðiÞ (i ¼ 1, 2) the time delay of the
    ð6Þ
>
> 0 0 0 0 target and jamming signal, respectively, and di(t)
>
: B1 ¼ , B2 ¼ :
0 1 1 0 (i ¼ 1, 2) the correlated complex Gaussian disturbance
with zero mean and covariance matrix d2 M:
Considering a single antenna radar system, a second-
order temporal pulse diversity scheme is applied.
Suppose radar transmission pulses in two consecutive
PRI are given by 2.3. Concept of rejection
 T In order to reject the false target, we resort to the
sT ¼ sT1 ðtÞ sT2 ðtÞ , ð7Þ second-order STBC here, i.e., inserting Equation (6)
Journal of the Chinese Institute of Engineers 4073

and N ¼ Nt ¼ M ¼ 2 to Equation (5), to obtain the 3. Simulations


transmission pulses given by The performance assessment of the presented algo-
8   rithm is generally indicated by the signal-to-jammer-
< s1 ðtÞ ¼ qT ðtÞ qT ðtÞ T , plus-disturbance ratio (SJDR) improvement factor
1 2
ð11Þ
: s ðtÞ ¼  H H
T
:
(IF), which is defined by
2 q 2 ð tÞ q 1 ð t Þ
SJDRout
In view of the fact that the WMF is inherent in a IF ¼ , ð15Þ
SJDRin
typical coherent radar system, we use it to filter the
received data in the second PRI. More precisely, the where SJDRin and SJDRout denote the input and
corresponding filter output is given by output SJDR of the WMF, respectively. Since closed
form expression for the operator T is not available, we
F½z2 ðtÞ ¼ sH 1
2 ðtÞM z2 ðtÞ extensively resort to Monte–Carlo techniques.
( In simulations, an LFM signal with bandwidth
X
H
¼ s2 ðtÞM 1
m ð2Þs2 ½t  m ð2Þ 20 MHz and time width 30 ms is utilized as the original
m radar signal. The first half of the vector is defined as
) q1(t) and q2(t) ¼ q1(t)ej. We utilize the model
X   
þ 0n ð2ÞT s1 t  n0 ð2Þ þ d2 ðtÞ fMgi,j ¼ d2 jij j as concerning the disturbance covari-
n ance matrix and we would investigate the two cases of
X   ¼ 0 (Gaussian white disturbance) and  ¼ 0.9
¼ m ð2ÞF s2 ½t  m ð2Þ (Gaussian correlated disturbance).
m
X    First, it is assumed that the target echo is received
þ 0n ð2ÞsH 1 0
2 ðtÞM T s1 t  n ð2Þ in the presence of a repeat false target. The effect of
n jamming on the performance of pulse compression is
sH 1 shown in Figure 1(a), where the jammer-to-signal ratio
þ 2 ðtÞM d2 ðtÞ: ð12Þ
P (JSR) is 6 dB. On the contrary, the output using the
Denote by m m ð2ÞFfs2 ½t  m ð2Þg the matched-filter- presented method is shown in Figure 1(b). It is clear
ing output that contains desired target echo.
P 0Evidently, that a deep fade is positioned at the peak location with
the second term in Equation (12), i.e., n ð2ÞsH
2 ðtÞ  respect to the jammer. In Figure 2(a), the IF of the
1 n
M Tfs1 ½t  n0 ð2Þg, relates to the jammer. Inserting presented method is plotted vs. the signal-to-
Equation (11) in the jamming term, if the transmission disturbance ratio (SDR) for  ¼ 0, and the associated
power of the radar stays unchanged during transmit- time-delay interval between the target and jammer is
ting within one PRI, we obtain 5 ms. The curves show that the method insures an above
X    23.5 dB performance gain with respect to the condi-
0n ð2ÞsH 1 0
2 ðtÞM T s1 t  n ð2Þ tions in which the JSR is set to 0, 6, and 12 dB. With
n
"    # the increase of the SDR, the IF stays about 28.2 dB.
X   T q1 t  n0 ð2Þ Another performance behavior is shown in
¼ 0n ð2Þ qT2 ðtÞ qT1 ðtÞ M 1
   ¼ 0: Figure 2(b), which assumes the same simulation
n T q2 t  n0 ð2Þ
parameters of Figure 2(a), but for the value of 
ð13Þ which is set to 0.9. It can be demonstrated that there is
Substitution of Equation (13) into Equation (12) yields a reduction in the performance gain of the presented
method with increasing SDR, and the IF remains
X 
F½z2 ðtÞ ¼ m ð2ÞF s2 ½t  m ð2Þ þ sH 1 around 22.4 dB. It is evident that, in the considered
2 ðtÞM d2 ðtÞ:
m ECM environment, the presented method performs
ð14Þ better in white Gaussian disturbance. The case of
jammers with different time delays with respect to the
It is clear that the jamming term has been eliminated target is analyzed in Figure 2(c) and (d). Here, the step
in Equation (14) and the target echoes are preserved. size is set to 0.2 ms and the SDR is fixed at 10 dB.
It follows that, resulting from the orthogonality of the Moreover, Figure 2(c) assumes  ¼ 0, while Figure 2(d)
codes used in pulse diversity, the performance considers  ¼ 0.9. The figures highlight that, when a
improvement is achieved from both the matched- repeat jammer is present, decreasing time delay does
filtering gain and the diversity gain, while not lead to obvious performance loss of the method.
the jamming signals are canceled out in the same This performance trend is reasonable because no
process. special time-delay constraints have been imposed
4408 G. Lu and B. Tang

(a) 0

Amplitude (dB)
−20

−40

−60
30 32 34 36 38 40 42 44 46 48 50
Time (µs)
(b) 0
Amplitude (dB)

−20

−40

−60
30 32 34 36 38 40 42 44 46 48 50
Time (µs)

Figure 1. Comparison of pulse compression results for one false target: (a) output of direct matched-filtering and (b) output
using this method.

(a) 29 (b) 29
JSR=0dB
27 27 JSR=6dB
JSR=12dB
IF(dB)

IF(dB)

25 25
JSR=0dB
23 JSR=6dB 23
JSR=12dB
21 21
−5 0 5 10 15 20 25 30 −5 0 5 10 15 20 25 30
SDR(dB) SDR(dB)
(c) 29 (d) 29
JSR=0dB
27 27 JSR=6dB
JSR=12dB
IF(dB)

IF(dB)

25 25
JSR=0dB
23 JSR=6dB 23
JSR=12dB
21 21
0 1 2 3 4 5 0 1 2 3 4 5
Time−delay(µs) Time−delay(µs)

Figure 2. SJDR IF curves for one false target: (a) IF vs. SDR,  ¼ 0; (b) IF vs. SDR,  ¼ 0.9; (c) IF vs. time delay,  ¼ 0; and (d)
IF vs. time delay,  ¼ 0.9.

upon the derivation of Equation (14). Indeed, this the jammer covers the target pulse entirely, which is
performance shows a favorable feature when the target shown in Figure 2(c) and (d).
and jamming pulses are overlapped. Particularly, the Second, two types of multi-false targets are ana-
method insures satisfactory performance levels even if lyzed in Figures 3 and 4. Type I is produced by
Journal of the Chinese Institute of Engineers 4095

(a) 0 (b) 0

Amplitude (dB)

Amplitude (dB)
−20 −20

−40 −40

−60 −60
35 40 45 35 40 45
Time (µs) Time (µs)
(c) 0 (d) 0
Amplitude (dB)

Amplitude (dB)
−20 −20

−40 −40

−60 −60
35 40 45 35 40 45
Time (µs) Time (µs)

Figure 3. Comparison of pulse compression results for multi-false target: (a) multi-false target type I; (b) multi-false target
type II; (c) rejection output of type I; and (d) rejection output of type II.

(a) 24 (b) 24
JSR=0dB JSR=0dB
23 22
JSR=6dB JSR=6dB
22 JSR=12dB JSR=12dB
20
IF(dB)

IF(dB)

21
18
20
19 16

18 14
−5 0 5 10 15 20 25 30 −5 0 5 10 15 20 25 30
SDR(dB) SDR(dB)

(c) 29 (d) 26
JSR=0dB JSR=0dB
27 JSR=6dB 24 JSR=6dB
JSR=12dB JSR=12dB
IF(dB)

IF(dB)

25 22

23 20

21 18
−5 0 5 10 15 20 25 30 −5 0 5 10 15 20 25 30
SDR(dB) SDR(dB)

Figure 4. SJDR IF curves for multi-false target: (a) type I : IF vs. SDR,  ¼ 0; (b) type I : IF vs. SDR,  ¼ 0.9; (c) type II : IF vs.
SDR,  ¼ 0; and (d) type II : IF vs. SDR,  ¼ 0.9.

repeating the stored radar signal in DRFM five times. LFM signal are placed in the adjacent vacant slots, and
Type II is also referred to as chopping and interleaving each selected segment is placed four times. The
(Sparrow and Cikalo 2006). Here, we assume five comparison of pulse compression results for type I
uniformly spaced sampled segments of the original jammer is exhibited in Figure 3(a) and (c) under the
6410 G. Lu and B. Tang

condition of JSR ¼ 6 dB and SDR ¼ 10 dB, and similar si the radar transmission signal
performance behavior for type II is shown in vector in the ith PRI
Figure 3(b) and (d). The effects of SDR on the T a linear operator
performance of the method for type I and type II are z the complex vector measured by
presented in Figure 4(a) ( ¼ 0) and (b) ( ¼ 0.9) and radars
Figure 4(c) ( ¼ 0) and (d) ( ¼ 0.9), respectively. m the complex amplitude of the
target signal
0n the complex amplitude of the jam-
4. Conclusions ming signal
c2 the variance of c
In this article, we have considered the problem of
d2 the variance of d
rejecting repeat jammers in white or correlated
n2 the variance of n
Gaussian disturbance with a joint fast/slow-time
m the time delay of the target signal
pulse diversity method. The proposed algorithm first
n0 the time delay of the jamming
separates the original transmission vector into two sub- signal
vectors. Then, using either or both of the sub-vectors
or their transformations to form an STBC signal, it
produces new sequences and transmits two sub-codes References
in one slow-time. In contrast to the slow-time method,
the presented method is practicable in rapidly varying Akhtar, J., 2009. Orthogonal block coded ECCM schemes
against repeat radar jammers. IEEE transactions on
conditions since no cancelation is required between
aerospace and electronic systems, 45 (3), 1218–1226.
two PRIs. With respect to the deception jamming
Alamouti, S.M., 1998. A simple transmit diversity technique
scenarios, the method also insures satisfactory perfor- for wireless communications. IEEE journal on selected
mance levels even if the target and jamming pulses are areas in communications, 16 (8), 1451–1458.
overlapped. Berger, S.D., 2003. Digital radio frequency memory linear
range gate stealer spectrum. IEEE transactions on aero-
space and electronic systems, 39 (2), 725–735.
Acknowledgments Gini, F. and Farina, A., 2002. Vector subspace detection in
compound-Gaussian clutter, part I: survey and new
The authors gratefully acknowledge the constructive com- results. IEEE transactions on aerospace and electronic
ments and suggestions of the reviewers.
systems, 38 (4), 1295–1311.
Lu, G., et al., 2010. Cancellation of complicated DRFM
range false targets via temporal pulse diversity. Progress in
Nomenclature electromagnetics research C, 16, 69–84.
Pace, P.E., et al., 2002. Digital false-target image synthesiser
Am, Bm the matrices used to design Q
for countering ISAR. IEE proceedings – radar, sonar and
c the correlated Gaussian clutter
navigation, 149 (5), 248–257.
vector Soumekh, M., 2006. SAR-ECCM using phase-perturbed
d the disturbance vector LFM chirp signals and DRFM repeat jammer penaliza-
I the identity matrix tion. IEEE transactions on aerospace and electronic
j the jamming vector systems, 42 (1), 191–205.
K the number of samples during a Sparrow, M.J. and Cikalo, J., 2006. ECM techniques to
single pulse counter pulse compression radar. United States Patent
L a lower triangular matrix 7,081,846.
M the normalized disturbance covari- Tarokh, V., et al., 1999a. Combined array processing and
ance matrix space-time coding. IEEE transactions on information
theory, 45 (4), 1121–1128.
n the white thermal noise vector
Tarokh, V., Jafarkhani, H., and Calderbank, A.R., 1999b.
p the complex amplitude of s
Space-time block coding for wireless communications:
Q the STBC matrix performance results. IEEE journal on selected areas in
qm the transmission signal set communications, 17 (3), 451–460.
R the normalized clutter covariance Xu, C. and Kwak, K.S., 2005. On decoding algorithm and
matrix performance of space-time block codes. IEEE transactions
s the radar signal vector on wireless communications, 4 (3), 825–829.

Das könnte Ihnen auch gefallen