Electronic Countermeasures (ECM) - MATLAB Simulations for Radar Systems Design - page 486

clutter). This design effectively reduces the amount of clutter that competes
with an individual target scatter by a factor of more than 40 dB, thus reducing
the C chaff ⁄ S by this same amount.

For extreme chaff cases where the initial WB range-Doppler image S/C is
negative, an N-pulse coherent sliding window routine can be applied to the
data prior to implementing the M 2 algorithm. For example, a 16 pulse coherent
sliding window can provide up to 12 dB of improvement. One
should ensure that the number of pulses integrated is less than the coherency
time of the target and clutter. Other constraints in implementing this approach
are to ensure that the target phase does not deviate very much during the inte-
gration period (to ensure optimum coherent processing gain) and the target
position does not migrate to another range and/or Doppler cell (often referred
to as range-Doppler walk). The zero Doppler filter (and/or near zero Doppler
filters) can be used to perform statistics on the clutter and to adaptively adjust
the optimal threshold setting to obtain low false alarms and high probabilities
of detection over time.
A model for the M 2 signal processor has been developed using MATLAB.
Fig. 10.10 shows a plot of the amplitude versus range and Doppler (256x256
range-Doppler image) of three constant -20 dBsm target scatterers that are
embedded in approximately -15 dBsm Gaussian white noise. In this figure, the
noise completely envelops the signal. These modeling results are comparable
to the output of a typical range Doppler imaging radar. Fig. 10.11 shows the
results obtained by executing the first two blocks of the M 2 signal processor.
As expected, the three scatterers rise from above the noise and now have an
ratio of approximately 7 dB.

SC chaff ⁄

SC chaff ⁄

Finally, Fig. 10.12 shows the results obtained by implementing the entire top
portion of the M 2 signal processing chain. No attempt was made to optimize
the threshold level. Instead, the threshold was manually set to -43 dB to allow
for some of the higher false alarms to be seen in the figure. The largest ampli-
tude false alarms are approximately -34 dB. Meanwhile, the amplitudes of the
target returns have been reduced (less than 1 dB) from that of Fig. 10.11.
Therefore, the improvement in Fig. 10.12 over that shown in Fig.
10.11 is approximately 8 to 9 dB. Hence, the processing gain attributed to the
M 2 signal processor is more than 20 dB above that of traditional range Doppler
processing.
In summary, one concludes that the M 2 signal processing algorithm for
detecting and tracking ballistic missile targets in highly cluttered environments
can provide better than 20 dB improvement over that of traditional
range Doppler processing techniques alone.

SC chaff ⁄

SC chaff ⁄