Advanced Technology on Radar Signal Processing

Wu Shunjun, Liao Guisheng and Bao Zheng

National Lab of Radar Signal Processing, Shannxi, Xi’an 710071, China

1 Introduction In known signal waveform case, according to the

theory of the matched filter, we can greatly improve the
The task of radar signal processing is the extraction
SNR and the doppler resolution performance by
of the target information from the radar echo. The function
coherently integrating (matched filtering) the target
of early radar was only the detection of targets from noise.
retums in a long time. But in practice the target returns are
Therefore, the signal processing function was very simple.
Following the urgent requirements with the radar, various usually a polluted version of transmitted signals, and
radars are developed. The first is surveillance radar. Then present nonstationarity in maneuver case and scatter
the tracking radar, phased array radar, airborne radar, migration through resolution cells (MTRC), in addition
multi-sites radar, imaging radar, HF OTH radar etc. come there exists instantaneous interference..Thus, for long time
on stage in succession. The functions of radar extend from coherent integration, we should develop efficient
the target detection and the coordinate measurement algorithms to mitigation the phase pollution of the signal
(range, angle of azimuth and elevation) to the in nonstationary case, to align signal envelope and to
measurement of doppler frequency, the judgement of compensate the high order phase of MTRC caused by
target property, anti-jammer anti-deception and anti-ARM targets’ maneuver. We also adopt proper system
and So on. This means that the modem radar should work configures to implement the long time coherent
in wide-band or multi-band, look every where all the time, integration. For example, we can adopt the phased array
possess multi-functions .... Then they must use advanced antenna, which transmits by a broad beam and receives by
signal processing technology to improve their capability a narrow beam, or transmits and receives by the multiple
of catch the target information. The advanced technologies beams simultaneously.
of radar signal processing involve the advanced The technology of the long time non-coherent
processing algolithms and advanced digital processing integration (such as video integration) to the output of the
hardware and software. matched filter can be used before detection to further
improve the detection performance. Obviously, decreasing
the detection threshold can increase the probability of
2 Algorithm
target detection at cost that the false alarm probability is
2.1 Signal processing algorithmfor HF/VHF radar increased unfortunately, which is prohibited in practice.
To detect small RCS targets such as stealth airplanes But when the trajectories of the target are strictly limited ,
and cruise missiles, an HFNHF radar can be used because decreasing the first detection threshold and eliminating the
it is appropriate to anti-stealth and Over The Horizon surplus false alarms in the duration of tracking allows to
(OTH) with a longer wavelength. The main problems to enhance detection performance. This is so called. Track
be solved include target detection in strong clutter Before Detect (TBD) technology. Considering the long
background, multipath, nonstationary signal and so on. We distance movement of targets we can adopt the Hough
must study efficient algorithms to improve the detection Transform to implement the long time non-coherent
and tracking performance of small RCS targets. integration.

l/$lO.OO 02001 IEEE.

0-7803-7000-7/0

16.

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on June 12,2023 at 13:19:59 UTC from IEEE Xplore. Restrictions apply.
 Bi-/multi-bases radar stations have shown their 2D space-time adaptive signal processing (STAP)
prominence in detecting stealth targets. Recently it is involves a large computational complexity. In the practical
proposed to utilize the broadcast or television signals and application we should try to decrease the degree of
mobile telephone stations to detect stealth targets and freedom in both the space and time domain. Only time
cruise missiles. This kind of passive radar does not domain processing of STAP based on a small quantity of
transmit the electromagnetic wave but only receives the spatial channels can improve more greatly the
reflectionhefraction wave from other transmitters. performance than on the single spatial channel, which has
Obviously it has many advantages. But detection and interested many people. A efficient method to decrease the
localization of targets from the clutter returns received by degree of freedom of the time domain is to implement
the passive radar is challenge. We must propose efficient doppler filtering preprocessing for the coherent pulse train,
algorithms to suppress clutter returns, strong direct wave and only extract the outputs of several doppler channels to
and multipath signals from noncooperation signal carry out STAP.
waveform. At present, the theory of STAP has been perfect on the
In addition, we still develop efficient algorithms to whole. But still some problems require further study,
improve doppler angle and range resolution of HFNHF which include the influence of the array of offset angle
OTH radar. and non-uniformity of the clutter ,and the extended
application of STAP such as STAP of over the horizon
2.2 Space-time signal processing algorithm for phased radar on the mobile platform on the sea (ground) and
array radar detecting and tracking the slow speea target, and so on.
The beam of phased array radar can be controlled Space-time coding transmitting is that the angular
flexibly. A important merit of digital beam forming (DBF) direction is coded on the transmitting array and is decoded
based on adaptive array processing is that it can suppress on the receiving array. The advantage is that the angular
interference, even that fiom the main lobe. As mentioned directions in illuminated sector are separated both on
above, the phased array radar can be adaptive a DBF transmission (by the coding ) and on reception (by digital
which can transmit by a broad beam and receive by beam forming), which possesses advantages when
multiple narrow beams at the same time, and allow to suppress the sidelobe clutter returns from airborne radar
suppress interference and implement long time coherent and counter fraudulent interference received by the
integration of targets simultaneously. sidelobe.
Combining the time information and the space
information to carry through the 2D space-time processing 2.3 Adaptive target recognition (ATR)
has been a hot research topic in recent years. Qne of the most important progresses that have been
Due to the doppler spreading of the clutter returns made in radar techniques is the increase in their resolution,
received by mobile platform, the suppression of clutter which has made it possible for the radar to perform not
returns is in essence 2D space-time filtering. In general, only the classical tasks of detection, location, searching
given the coupling between the doppler frequency of the and tracing, but many advanced functions including target
clutter returns and their arrival angles, by the 2D doppler- imaging and recognition . In recent years, high resolution
angle adaptive processing , we can efficiently suppress the radar target recognition has received much intensive
clutter and improve the detection performance of the attention from the radar as well as signal processing
targets, especially can detect the targets moving slowly on community and has obtained great progress on many
the ground. This is an important technology used on the essential problems, but some problems still do not be
high speed aircraft or the pilotless aircraft to monitor the solved well. First, the aspect-smoothing of range profiles
ground. has remained for many years the most challenging task in

17

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on June 12,2023 at 13:19:59 UTC from IEEE Xplore. Restrictions apply.
 range profiles recognition. The algorithm of the aspect- Moving Target Indication) makes enormous progress of
smoothing of range profiles can suppress scatter cross the detection performance of radar target in clutters. After
terms in the bispectral domain by utilizing their statistical that, the digital processing has permeated every
properties, and reconstruct scatter auto terms from subsystems. The technology of signal processing is
accumulation of radial slices of the smoothed bispectra. changing with each passing day.
Second, the extraction of shift invariant features of range The digitization frequency of radar signal is moving
profiles is another important topic on target recognition. up to the front of radar receiver. Usually, the
The circularly integrated bispectra is a kind of eficient digitization of signal is made in zero intermediate
shift invariant features of range profiles. frequency (IF). Now it is moving up to low IF and
Recently important progresses have been made on the more high frequency. This kind of change benefit
smart processing of ATR based on neural network, mainly from the progress of the A/D converter.
wavelet transform and fuzzy theory. The digital receiver makes the unification of the
multi-channel receivers easy. The digital receiver has
2.4 Radar imaging algorithm a large dynamic range region of signal. These will be
With the broadband signal used widely, the radar benefit to the adaptive digital beam forming and
imaging, including SARJISAR imaging, has received space-time adaptive signal processing.
intensive attention. In order to obtain high-resolution radar The progress of D/A performance causes the quick
images, it is necessary to maintain the precise phase development of the DDS (Direct Digital Synthesis)
relationship between return signals. This requires precise technology. It improves the signal quality and
motion compensation, because the unplanned frequency agility capability of the radar transmitter.
perturbations of the relative motion between the radar and Then the SCV and ECCM capability will be
the scatters causes phase errors, which may destroy the enhanced.
phase coherence of return signals and blur radar images The strong computational capability changes the
severely. Motion compensation can be done by two hardware of radar signal processor.
separate processing approaches: range alignment and The digital processing of signal in a modem radar
initial phase error correction. needs more and more computational requirement. The
At present, important progresses have been made on STAP, SAR, ISAR, SIARf2] and ATR all need the
the research of precise 3D SAR imaging, detection of massive processing volume. For this kind of high
moving targets of SAR and the ISAR imaging of speed real-time processing, the computational amount
maneuvering targets. We can obtain 3D images by single will be greater than tens GFLOPS (10 floating point
'
pulse interference SAR imaging and measuring the height. operations per seed). We can not use single DSP
Based on the high-order ( 2 2 orders) phase polynomial (digital signal processor) to accomplish the real-time
signal model of the return signals, we can achieve ISAR processing.
dynamic images of the maneuvering targets. To solve the problem, the optimization of radar signal
processing is researched, and the parallel processing
technology of Multi-DSPs is developed alsof3].The
3 Signal Processing Hardware and Fig.1 shows a structure of the modern radar signal
processor. The radar signal of low IF is changed to
Software data in the A/D converts. All the functions the radar
signal processing are completed in the parallel
The application of digital technology is a revolution
processor. These processing functions, for instance,
in the radar history. It begun in 1970s. DMTI (Digital
pulse compression, channel equalization, MTI (or

18

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on June 12,2023 at 13:19:59 UTC from IEEE Xplore. Restrictions apply.
 MTD.PD), digital beam forming, CFAR detection and The radar Engineer can use the radar signal
so on, are all accomplished by programming. modelling and simulation library (RSPMSL)[41 to

-
construct the soft-system in the computer. Then they
1
can simulate and optimize the performance of a radar
signal processing system. The RSPMSL can help this

._ A/D * Parallel
_7/ Processor
--+
+ I/O
to avoid complex programming. The design process
will be speeded.

__*

Reference

Clock Timer And Controller M.B.Skolnik, etc. “The Engineering of System

Engineers”, IEEE AES Magazine, Jubilee Issue,
oct., 2000.
Fig. 1 a structure of the modern radar signal processor
Wu Jianqi, etc., “The Research and experiment on
The structure of the parallel processor is the spare-array synthetic Impulse and Aperture
changeable too by programming and the arrangement Radar”, Proceedings of the CIE 7‘h National
of links between the DSPs. Radar Conference, Nanjing, 1999
The modelling and simulaton of radar signal
processor speed up the update of signal processing So Tao and Bao Zheng, “VLSI Parallelor in
design technology. Space-Time Adaptive Signal Processing”,
Proceedings of the CIE 7‘h National Radar
Advanced radar require excellent functions and Conference. Nanjing, 1999
performances. Following the quick development of
computer technology and the idea of Electronic Wu Shunjun, Zhang Lingrang, Su Tao and Li
Design Automation (EDA). The modelling and Ming, “United Method of Modelling Simulation
simulation of radar signal processor is proposed and and Design for Radar Signal Processing systems”,
developed quickly. Fig.2 shows a united design The Proceedings of the International Conference
method for radar signal processing system. on Radar Systems, France, May 17-21, 1999

System Function & Specifjcation

+
t System Construction
c
4-

+
Modelling
Signal Processing Algorithm ‘z:’
Simulation +
System Simulation 8 Evaluation
1 .L

Division of Circuit

Design Circuit Simulation 8 Optimization

+ ~ ~~~

I I Prototype of System
I
Fig.2 United Design Method

19

Authorized licensed use limited to: INDIAN INSTITUTE OF TECHNOLOGY MADRAS. Downloaded on June 12,2023 at 13:19:59 UTC from IEEE Xplore. Restrictions apply.