Modern radars can use a wide variety of waveforms with varying performance characteristics. In general a waveform can be tailored to achieve good Doppler or good range resolution, but not both simultaneously. This is a problem in heavy clutter environments, typified by an airborne radar seeking to detect slow moving ground targets. Detection in this situation is made difficult by the relatively low Doppler component of the target returns, making them almost indistinguishable from the Doppler spread of the clutter.
A number of possible approaches to improving radar performance in this type of environment are available. For example, the use of a broadband transmitter permits a high degree of tailoring of radar capability via the use of multiple waveforms. Clutter mapping and optimal scheduling of the waveforms also permit tailoring of waveforms to best match the working environment of the radar. Complementary waveforms are one multi-waveform approach to optimization of performance via waveform design, although they have not yet received serious consideration for a practical system. These various techniques, alone or in combination, also offer the possibility of adaptive adjustment of waveforms to optimize radar performance. However, because of the high data rates and rates of change of the radar scene, manual optimization of the performance of a modern radar by pulse tailoring is not possible. Current adaptive schemes are founded on signal-based criteria such as maximizing detection probability, but do not attempt to use information available in the radar return signal about the environment of the radar system.
Our work has focused on the use of a probabilistic data association tracker that applies all information to achieve detection and tracking over a variety of radar environments and the adaptive and optimal tailoring of complementary and other waveforms, based on the signal and clutter environment of the radar and on scheduling of waveforms. The latter process is expected to achieve much higher performance through optimized ambiguity functions.
Our investigations have established that the techniques have considerable potential, lowering thresholds of detection by several dB. We are now refining the theoretical basis for the techniques and developing a simulator that allows evaluation of the performance of the techniques on radars operating in a variety of scenarios and environments.
