Fast ML-Assisted Interference Estimation and Suppression for Digital Phased Array Radar

Radars must operate in environments where interference can potentially degrade performance. With a fully digital array, classical adaptive radar signal processing methods suffer from higher order of computations and latency overhead. This is a challenging problem in the era of dynamic spectrum sharing where the interference sources can be highly variable. In this paper, we propose a fast ML-assisted method for detection of interference sources and estimation of their directions of arrival. These estimates are used for computation of adaptive weights with a covariance-based method that reduces the number of complex matrix inversion operations. Using simulation-based evaluations, we compare the results of our proposed approach with the classical methods.