A Gaussian particle filter for track-before-detect

A novel track-before-detect algorithm is proposed to perform detection and tracking a weak target in radar measurements. Firstly, a Gaussian particle filter based track before detect is presented for tracking target. Secondly because of the Gaussian particle filter is only tracking algorithm and can not perform detection target, the likelihood ratio based the Gaussian particle filter is derived. Finally based on this deduction, sequential fixed size (FSS) likelihood ratio test (LRT) is used to detect target. Simulation results show that this novel algorithm can realized detection and tracking a weak target simultaneously, and the performance of detection and tracking is improved than the particle filter.