Aerospace and Electronic Systems Magazine May 2017 - 9

Liu et al.

Figure 9.

Signal processing diagram.

technology provides high computational power, making it an ideal
solution for DDC and downsampling processing in the data acquisition board. The usage of DSP gives more freedom in the implementation of different signal processing algorithms in the signal
processing board. Figure 8 shows a detailed image of the radar
front end and the digital section.

RADAR DSP
Processing of the digital baseband signal for detection and localization of targets is accomplished in the signal processing board.
According to (3) and (4), beat frequency fBeat of a single target is a
signal with single frequency in each receiving pulse, and the beat
frequency can be obtained with the fast Fourier transform (FFT).
The phase difference between two adjacent pulses is caused by
the Doppler frequency fd only; the Doppler frequency can also be
achieved with FFT.
The whole processing procedure is illustrated in Figure 9. A
Hanning window is applied to each digital baseband signal pulse
for the reduction of sidelobes in the spectrum. The FFT method
is then applied to each pulse to obtain the beat frequency spectrum, i.e., the range spectrum. After that, another FFT processing
is adopted for the obtained range spectrums that are arranged in sequence for the Doppler frequency spectrum. As a result of the two
MAY 2017

dimensions of FFT processing, the range-Doppler processing is
accomplished and the range and Doppler map is finally achieved.
Targets with different range and different radial velocity are separated in corresponding range and Doppler resolution cells. According to the theory of LFMCW radars, the time delay of the received
echo can be obtained by simply measuring the frequency fr, thus
calculating the range R between the targets and the radar facility
from R = cfr /(2k). The relative radial velocity of a detected target
can be calculated from v = cfd/(2f0) with the Doppler frequency fd.
With the range and Doppler map, a cell-averaging constant
false alarm rate (CA-CFAR) detector [28] is employed to pick up
targets of interest in different range-Doppler resolution cells. When
targets of interest have been detected, along with their ranges and
radial velocities, the remaining task is to obtain their DOA estimations. Since we have made most of the width of the train for the
receive array, system angular resolution of 0.36° may be obtained.
The system may have sufficient angular accuracy in a single target scenario when the signal-to-noise ratio (SNR) is high enough.
However, there are always many closely spaced targets such as catenary poles in the radar observation area. With such limited system
angular resolution, it is not possible to separate two closely spaced
targets 3 km away with the conventional DBF method.
The superresolution algorithm MUSIC [14], [15], [29], [30]
is utilized here for its better performance in DOA estimation.

IEEE A&E SYSTEMS MAGAZINE

Table of Contents for the Digital Edition of Aerospace and Electronic Systems Magazine May 2017