In this paper, we focus on the performance improvement of dynamic programming based track-before-detect (DP-TBD) in heterogeneous Rayleigh background, in which case the surveillance scene is divided into two parts with separate independent and identically distributed Rayleigh probability density functions. Original DP-TBD (DP-TBD using signal amplitude scoring function) suffers significant performance loss under this condition. Optimal DP-TBD (DP-TBD using log-likelihood ratio (LLR) scoring function) usually can not be obtained since the clutter edge position and clutter statistics are unknown. To overcome these challenges, we propose an improved DP-TBD strategy, which applies a data pre-processing step to the measurement frame before DP processing, the clutter is then suppressed. Simulation results show that the proposed DP-TBD outperforms the original DP-TBD, the false tracks is reduced over 80%, and the performance loss is less than 0.5dB in comparison with the optimal DP-TBD.
展开▼