Subspace Detection for Range-Spread Target to Suppress Interference: Exploiting Persymmetry in Non-Homogeneous Scenario

Abstract

This paper deals with subspace detection for rangespread target in non-homogeneous clutter with unknown covariance matrix where structured interference is presented in the received data. Through exploiting the persymmetry of the clutter covariance matrix, we propose two adaptive target detectors, which are referred to as persymmetric subspace Rao to suppress interference and persymmetric subspace Wald to suppress interference ( "PS-Rao-I " and "PS-Wald-I "), respectively. The persymmetry-based design brings in the advantage of easy implementation for small training sample support. The signal flow analysis of the two detectors shows that the PS-Rao-I rejects interference and integrates signals successively through separated matrix projection, while the PS-Wald-I jointly achieves interference elimination and signal combination via oblique projection. In addition, both detectors are shown to be constant false alarm rate detectors, significantly improving the detection performance with other competing detectors under the condition of limited training.

A novel two-step detection algorithm for range spread target

“…Although many one pulse-group detectors have been developed in recent years, they can effectively improve the fusion gain of RST cells at different distances [3][4][5] . In contemporary times, the field of object detection has witnessed the advent of several nascent machine learning algorithms.…”