Radar code optimization for moving target detection

In this paper, we study the problem of optimal pulsed-radar transmit code design for detection of moving targets in the presence of clutter. To this end, a new data modeling is introduced and optimal detectors for known and unknown target Doppler shift are presented. In the case of unknown Doppler shift, the expressions of the optimal detector and its performance metrics are too complicated to be amenable for code design. Therefore, we consider a substitute metric whose maximization results in maximization of a lower bound on the J-divergence. For a known target Doppler shift, the metric directly determines the performance of the optimal detector. In order to solve the emerging highly non-convex code design problem, we devise a computational framework (which we call CADCODE) that lays the ground for an optimal code design via a cyclic minimization. As low peak-to-average-power ratio (PAR) codes are of interest in many applications, we apply the CADCODE framework to obtain such codes. Several numerical examples are provided to illustrate the performance of the proposed algorithm.