A configurable compressive acquisition matrix for radar tracking using particle filtering

During the process of tracking using radar a set of delay-Doppler shifts along with their associated probabilities represent known information on the target state. This information can be used to offset the error introduced by compressive sensing and processing (CSP) which operates at a sub-Nyquist rate. In past work the CSP mechanism was adapted using the set of delay-Doppler shifts that may appear in the radar return waveform due to the presence of a single target. In this work a new method is proposed that utilizes not only the set of Delay-Doppler shifts, but also the probability associated with each delay-Doppler pair of appearing in the radar return waveform. Both theoretical and simulation based evidence is provided to demonstrate the improvement in estimation performance by the proposed method compared to previous CSP methods.