IJDI-ERET

IJDI-ERET INTERNATIONAL JOURNAL OF DARSHAN INSTITUTE ON

ENGINEERING RESEARCH & EMERGING TECHNOLOGIES
Vol. 2, No. 2, 2013
www.darshan.ac.in
(Research Article)
An Algorithm for the Interception and Analysis of Pulse
Compression Radar Signals by Digital Receiver
AK Singh1*, Dr. K. Subba Rao2
1
Scientist- Defence Electronics Resea rch Laboratory (DLRL), Hyderabad
2
Professor, ECE Department, Osmania University, Hyderabad

Abstract

Pulse Compression is an Electronic Protection (EP) technique being used in the modern radars to achieve the advantages of a
good range resolution and target detection capability at low power levels and simultaneously offering Low Probability of
Intercept (LPI) features. Pulse Compression Radars are being integrated into the many platforms and weapon systems of the
modern battlefield. The Detection and analysis of pulse compression radars demand ELINT Receiver with very high sensitivity
and capability to extract inter and intra pulse modulation parameters. In this paper, Digital Receiver, a state of the art single
board solution for the detection and analysis of modern radar signals in real time is presented. A novel algorithm based on the
All Phase FFT is being implemented successfully in the Digital Receiver Hardware in real time making the system useful for
both the ES/ELINT applications. The Algorithm is capable of extracting all the basic as well as advanced parameters of
frequency and phase modulations such as Chirp, Barker, and Poly-phase codes. The paper describes various pulse compression
radar waveforms and thoroughly discusses the proposed Digital Receiver Hardware and implementation of signal processing
algorithms for the interception and analysis of complex pulse compression radar signals. The simulated and measured test data
for various digitally modulated radar signals is presented. This type of receiver plays a vital role in the modern battlefield for
ES/ELINT application.

Keywords: LPI- Low Probability of Intercept, ELINT-Electronic Intelligence; ES-Electronic Support, FPGA- Field
Programmable Gate Array, EW-Electronic Warfare, apFFT-All Phase FFT

1. Introduction pulse compression filter to yield a narrow compressed pulse

response with a main lobe width of approximately 1/Bw that
Since the dawn of radar, both radars and electronic warfare does not depend on the duration of the transmitted pulse. The
(EW) systems evolved as state-of-the-art technologies three general types of LPI radar using pulse compression are
weredeveloped and adopted. This cat-and-mouse race the: (a) frequency modulating radar that includes FMCW and
continues to this very day, and its pace is even accelerating. chirp signals, (b) the phase modulating radar that includes
In the last 10 years emerging technologies pushed the Poly-phase modulation and (c) the radar that is a
systems to a new level of sophistication, offering ultimate combination of both (a) and (b)[2]. The interception and
performance [1]. Pulse Compression in radar signals is one analysis of pulse compression waveforms’ using both
such technology that has a dramatic influence of on the frequency and phase modulation is very much needed for the
electronic battlefield. modern EW system for effective performance on the
battlefield and is the subject of the paper.
Pulse Compression radar transmits a long pulse with pulse
width T and peak power Pt which is coded using frequency or 2. Digital Receiver Hardware
phase modulation to achieve a bandwidth Bw that is large
compared to that of an un-coded pulse with the same Keeping in view the complexity of the pulse compression
duration. The transmit pulse width is chosen to achieve the radar signals and the associated algorithms for the detection
single pulse transmit energy, that is required for target of pulse compression radar signals, Digital Receiver
detection or tracking. The received echo is processed using a hardware containing the state of the art data acquisition and
processing devices has been designed by adapting the latest
*
Corresponding Author: e-mail: singh.ak@dlrl.drdo.in, Tel-+91-040- PCB board design practices. The Digital Receiver hardware
24582632, Fax-+91-040-24582632 is shown in figure 1.The Digital ReceiverHardware will have
ISSN 2320-7590
two modules namely IF & Clock synthesizer section and
 2013 Darshan Institute of Engg. & Tech., All rights reserved Digital Receiver Main board. Digital Receiver hardware
 International Journal of Darshan Institute on Engineering Research and Emerging Technology
Vol. 2, No. 2, 2013, pp. 18-22

receives IF input in the frequency range of either 160±20 further the exact subtype within the modulation i.e. Barker,
MHz (140 to 180 MHz) or 1000 ±250MHz (750 to 1250 Frank, Poly-phase etc., can be determined.
MHz). The inputs are amplified and filtered to eliminate any
spurious signals with a gain of 12dB to meet the input power
level requirements of ADC which is present on the main
board.These signals are fed to ADC which operates at 1350
MHZ sampling rate.

Figure 2. N-Order FFT Spectrum Analysis

3.1 All Phase FFT: All Phase FFT [4] is an algorithm which
overcomes the disadvantages of conventional FFT based
algorithm and achieves the high accuracy phase measurement
Figure 1: Digital Receiver Hardware without the need for spectrum correcting operations. Also,
All Phase FFT is superior to conventional FFT in inhibiting
The sampled data will be fed to Virtex-6 SX-475 FPGA. the spectral leakage.
The FPGA receives the sampled data in real time for signal
processing using the advanced algorithm based on All Phase
FFTand to compute/process the signal parameters. The
Hardware has two TS-201 DSP’s for further processing. The
output of the FPGA is fed through a 51 pin ‘D’ Connector for
debug and also through VME connectors to ESM processor
card for further processing. The algorithm that is being
implemented inside the Digital Receiver Hardware is
discussed below.

3. Algorithm For Interception And Analysis

FFT is suitable for extracting the carrier frequency,

amplitude and also pulse parameters like Pulse Width, Time
of Arrival(TOA), but it has two drawbacks which limits its
use for phase measurement, they are firstly, only when Figure 3. All Phase FFT Spectrum Analysis
sampling frequency equals the integer times of FFT’s
frequency interval, phase value can be computed accurately. 3.1.1 Principle of All Phase FFT: The principle of All Phase
Otherwise the estimated phase will deviate from the real one, FFT is represented pictorially in the figure 3. The principle of
which is known as the case of ‘non synchronous sampling’. traditional FFT to differentiate the All Phase FFT from
To overcome this shortcoming, additional measure of normal FFT is shown in figure 2. An example of All Phase
correcting spectrum should be taken. Secondly, FFT’s FFT Spectrum Analysis is shown in figure 4.
inherent effects of spectral leakage will also reduce the
precision of phase measurement [3]. The input data is firstly windowed by a (2N-1) length
window wc, then every two samples with the space between
A custom Algorithm based on All Phase FFT has been N intervals (except the middle element x(0)) are summed up
developed and implemented to overcome the above difficulty. thus one new N-length data vector y=[ y(0), y(1), …, y (N-1)]
With this algorithm, all the advanced parameters of the Intra can be formed. Lastly, implementing DFT/FFT on the new
pulse modulated radar signal such as the sub code duration, vector will generate the outcome of apFFT.
bandwidth, phase change pattern, frequency change patterns
within the pulse, type of modulation i.e. frequency or phase

19
 International Journal of Darshan Institute on Engineering Research and Emerging Technology
Vol. 2, No. 2, 2013, pp. 18-22

, , , is selected as an
example to compare the traditional DFT phase spectrum with
apDFT phase spectrum.

The N-order normalized traditional windowed FFT

spectrum X(k) and the N-order normalized dual windowed
All Phase FFT spectrum Y(k) with can be respectively
represented as
(4)

(5)
Figure 4. All Phase FFT Spectrum Analysis-Example
Calculation
From Equation (4) and Equation (5), we can find that the
The implementation shown above is an optimum phase value of traditional FFT spectrum X(k) has a close
implementation which makes apFFT applicable and relationship with its counterpart frequency offset ‘(β-k) Δω’.
convenient to be implemented [5]. ‘wc’ introduced in Figure On the contrary, the phase value of apFFT spectrum Y(k) is
4 is named as ‘convolution window’, which equals the constantly equal to the initial phase �0, independent of the
convolution of one ‘front window f’ and one reversed "back frequency offset ‘(β-k)Δω’.Moreover, the ‘�0’ is exactly
window b ",that is equal to the theoretic phase value of the (2N-1)-length input
data vector’s central sample ‘x (0)’.In conclusion, having
w (n) = f (n) * b(n) (1) implemented apFFT on the (2N-1) input data, we can directly
obtain the central sample’s theoretic phase value from the
According to Equation (1) peak spectral line without any additional correcting
operations. Thus we say that apFFT possesses the property of
‘phase invariant’.
(2) 3.2 Simulation of All Phase FFT Algorithm: MATLAB
simulation has been carried out on sinusoidal signal to clearly
view the working principle of All Phase FFT and to compare
If RN represents the N-length rectangular window, then it against the No Window and Windowed FFT. The
apFFT can be classified in 3 categories: The case that simulation results for this have been shown in the Figures
satisfies f= b=RN is called no-windowed apFFT; The case 5(a) and 5(b).
that satisfies f or b = RN is called single windowed apFFT
and The case that satisfies f =b ≠ RN is called dual-windowed
apFFT. The FFT used for basic parameter measurement
above refer to the case of All Phase FFT with no window.

3.1.2. Amplitude Spectrum of APFFT: The amplitude of all

phase spectrum is as follows [6]

(3)

This is square of the traditional Fourier Transform

frequency spectrum amplitude. It can also be noted that all N
spectral lines amplitudes obey the same square law, which Figure 5(a). Transformations on the input signal in All Phase
means that the amplitude ratio of adjacent lines to the peak FFT Algorithm
line decreases more acutely according this square relation. As
a result, apFFT possesses a better quality than FFT in
inhibiting spectral leakage.

3.1.3 Phase Spectrum of All Phase FFT: A single-frequency

complex exponential sequence ,
20
 International Journal of Darshan Institute on Engineering Research and Emerging Technology
Vol. 2, No. 2, 2013, pp. 18-22

distinguish one LPI radar signal from another and they are
required for effective exploitation. After getting all the
parameters the signal is being classified according to the
database and the corresponding signal type and the basic as
well as advanced parameters are being sent as the Pulse
Descriptor Word (PDW).The algorithm flow is being shown
in figure 6.

5. Implementation and Testing

For testing the effectiveness of the algorithm against the

detection of various Intrapulse Modulation signals, the signal
being generated from the Arbitrary Waveform Generator [7]
Figure 5(b). Magnitude and Phase Spectrums of FFT, is digitized by the High Speed ADC and is being latched into
Windowed FFT and All Phase FFT the FPGA present in the Digital Receiver Hardware and is
being processed to extract the various basic and advanced
In Figure 5(a) the top leftmost plot shows the input signal, parameters of the Intrapulse modulated signals. The results
followed by Hamming and Hanning window signal plots and are monitored using the Chip scope Pro tool of FPGA. The
their signal transformations and the last one shows the signal test setup used for testing the algorithm using the Arbitrary
after windowing and adding as shown in figure 5 (a) for the Waveform Generator and Digital Receiver Hardware is
All Phase FFT. In Figure 5(b) the top leftmost plot shows shown in figure 7.
the magnitude and phase plots of FFT, followed by
magnitude and phase plots FFT with Hamming and Hanning
window.The bottom most ones show the phase and
magnitude plots for the All Phase FFT. From Figure 5(b) we
can clearly say that with the All Phase FFT the phase of the
signal can be identified clearly as the frequency and also the
phase is invariant for nearby indexes also and is completely
unambiguous.

4. Intrapulse Modulation Extraction:

Having simulated the All Phase FFT, this algorithm is

chosen as the base for the extraction of phase pattern from the
Intrapulse modulated phase modulated radar signals. After
the phase collection from the All Phase FFT the phase
information within the N no of samples (512 in our case) will
be obtained. This phase is again unwrapped to get the actual
phase. Whenever there is a phase change happens within the
samples under consideration the phase from the central bin
will deviate from the actual one. From the information
obtained through the previous frame and next frame the phase
change is calculated. If there is no phase change is present Figure 6.Basic Flow of Algorithm
then the signal is considered as a pulse modulated signal.
Similarly if the frequency from the successive frames differs
then the signal will be classified as a Frequency modulated
signal and the corresponding parameters like start frequency,
stop frequency and rate of change of frequency are being sent
as output. If the signal is having phase modulation then the
phase change pattern is obtained. Here the results obtained
using the processing will be used to identify the various
modulations based on the basic parameters of the radar such
as Frequency, Amplitude, Pulse Width, Time of Arrival,
Direction of Arrival and advanced parameters like
Modulation bandwidth,Modulation period, Code period and
Figure 7.Test Setup for Evaluation of Algorithm
time for an LPI Radar. These are the parameters that

21
 International Journal of Darshan Institute on Engineering Research and Emerging Technology
Vol. 2, No. 2, 2013, pp. 18-22

The types of the signals and the Intra pulse modulation The ELINT Digital Receiver presented in this paper is an
parameters that can be obtained with the above algorithm are advanced ES/ELINT receiver for the detection and analysis
shown in Table 1. of modernPulse Compression Radar signals. The Algorithm
embedded in the hardware is capable of extracting all the
Table 1: Effectiveness of Custom Algorithm basic as well as advanced parameters of the phase as well as
Parameter Design Goal Design frequency modulated Intrapulse Radar Signal. The system of
Achievements this kind which is able to extract the basic as well as
advanced parameters of the radar signal in real time plays a
Type of Chirp, Barker, Chirp, Barker, very crucial role in estimating the enemy radar signals and in
Signals Frank, Frank,P1-P4 effectively countering the associated threats.
P1-P4
References
Parameters Centre Centre Frequency,
of Chirp Frequency, Chirp Bandwidth, 1. Uri Barkan&Shuki Yehuda, “Trends in Radar and
Signals Chirp Chirp Rate Electronic Warfare Technologies and their Influence
Bandwidth, on the Electromagnetic Spectrum Evolution ”, 2012
Chirp Rate IEEE 27th Convention of Electrical and Electronics
Engineers in Israel, pages 5 2012
Phase Sub Code Sub Code 2. Philip E Pace, “Detecting And Classifying LPI
Modulation Duration, Phase Duration, Phase Radar”, Second Edition, Artech House, Inc.,
Parameters Pattern, Type of Pattern, Type of Norwood, Massachusetts, 2009.’
Phase Code Phase Code 3. “The Fundamentals of Signal Analysis".,
Application Note No 243,Agilent Technologies.
www.agilent.com/find
Table 2: Extracted Phase Patterns for Barker and Frank Code 4. WANG Zhaohua, HouZhengxin and Su Fei, “All
phase FFT spectrumanalysis, "Journal of China
Institute of Communications, Vol. 24, No. 11A, Nov
2003, pp. 16-19.
5. Huang Xiangdong and Wang Zhaohua, “Phase
Difference Correcting Spectrum Method Based on
All-phase Spectrum Analysis,” Journal of
Electronics & Information Technology, Vol. 30, No.
2, Feb 2008, pp. 293-297.
6. Guo Jing, Li Nan, Zhou Zhou, Yi Wei, XuXin,
Wang Yinan, “A Novel Digital Demodulation
Method for MFEIT based on APFFT”, International
Table 2 shows the output phase of a Barker and Frank code Conference on Intelligent System Design and
that are being extracted with the custom algorithm. It is also Engineering Application, 2010 IEEE,DOI
observed that the phase error is within ±5˚, which is sufficient 10.1109/ISDEA.2010.239.
to extract all the poly-phase modulations. 7. BeateHoehne, Germany , “Digital Up Conversion
VS IQ modulation using a wideband Arbitrary
6. Conclusion Waveform Generator”, 978-1-4673-0700-
0/12©2012 IEEE.

Biographical notes
A.K. Singh completed M.E. in Digital System (ECE) from Osmania University in 2003. He joined Defence Electronics Research Laboratory
(DLRL) in 1996. Currently he is Scientist-‘F’ and leading a team working on design & development of real time Digital Receiver. His area of
interest includes high speed board design, Time-frequency signal processing, and EW Receiver design.

Dr. K SubbaRaohas Graduated from S V University, Tirupati, India. He obtained the Master’s and Ph.D. degree from Osmania University
(O.U), Hyderabad, India. He Joined Osmania University in 1981 as lecturer and served as a headed the ECE department. He has published more
than 200 research papers in national and international Journals/ conferences. His current interests include signal processing for Radar & Spread
spectrum applications and Bio Medical signal processing.

22