Knowledge-based wideband radar target detection in the heterogeneous environment

In this paper, we address wideband radar target detection in the heterogeneous environment. Firstly, a linear model of the wideband radar target return with the steering vector dispersion is established. Secondly, the heterogeneous clutter is modeled as a two-dimensional wide-sense stationary (WSS) process with inverse complex Wishart distributed random covariance matrices in the time-space and frequency domain. Then, several generalized likelihood ratio test (GLRT) based detectors are designed, some of which integrate the prior knowledge of the clutter covariance matrix with the Bayesian approach, while the others are with the heuristic approach. Finally, the performance of the detectors is evaluated by simulations, and the results show that the detectors based on the Bayesian approach outperform the other detectors.