Page 367 - Fundamentals of Radar Signal Processing

motion is detected and compensated, the MTI filter null will not be centered on

the clutter spectrum and cancellation will be poor.
The largest source of clutter offset and spreading is radar platform motion.
Recall from Chap. 3, Eq. (3.5), that the motion-induced clutter bandwidth is

(5.72)

The offset in center frequency of the clutter spectrum can be as much as a few
kilohertz for fast aircraft when forward looking, while the motion-induced
spectral spread can be tens to a few hundreds of hertz for a sidelooking
configuration. This clutter spreading adds to the intrinsic spread of the clutter
spectrum due to internal motion and can often be the dominant effect determining
the observed clutter spectral width and therefore determining the MTI

performance limits.

5.3 Pulse Doppler Processing

Pulse Doppler processing is the second major class of Doppler processing.
Recall that in MTI processing the fast time-slow-time data matrix is highpass
filtered in the slow-time dimension, yielding a new fast-time/slow-time data
sequence in which the clutter components have been attenuated. Pulse Doppler
processing differs in that filtering in the slow-time domain is replaced by

explicit spectral analysis of the slow-time data for each range bin. Target
detection is then performed directly on the range-Doppler matrix of data.
Because the range-Doppler matrix is the fundamental data quantity in pulse-
Doppler processing, the first step is to form it from the fast-time/slow-time CPI
matrix by computing the one-dimensional spectrum of the slow-time signal in
each range bin. The spectral analysis is most commonly by far performed using
the fast Fourier transform (FFT) algorithm to compute the discrete Fourier
transform (DFT) as shown in Fig. 5.15, but other techniques can also be used.

The computed DFT is a frequency-sampled version of the DTFT of the slow-
time signal. The figure emphasizes the fact that the number of frequency samples
K does not have to equal the number of slow-time samples M; often K > M as
shown.