A bayesian approach for dual-knowledge-aided target detection and performance analysis in heterogeneous environments

Pucheng Jing, Yongchan Gao, Jun Liu, Lei Zuo, Zhiwen Xu

Abstract: In this paper, a Bayesian approach for dual-knowledge-aided target detection is designed in heterogeneous environments, where three novel detectors resorting to generalized likelihood ratio, Rao and Wald criteria via two-step procedure are proposed. Specifically, we derive the detection statistics with assuming known covariance matrix structure in heterogeneous environments under a Bayesian framework in the first step. In the second step, a dual-knowledge-aided covariance matrix structure is estimated using both information of the prior distribution and Centro-Hermitian structure of clutter and used as a substitute in the statistics. Then, a performance analysis regarding the proposed approach is developed from three perspectives, including the relationship of the three detectors by a unified expression, the computational complexity, and inaccurate information error modeling. Finally, numerical results project that the performance of the proposed approach over suitable counterparts.