Improved Residual Cumulative Resampling PF-PHD Filter Based on Stratified Rough Sorting

Resampling is the main step that constrains the hardware implementation of Particle Filter- Probability hypothesis density(PF-Phd). This step requires particle deletion and replication, which brings huge latency and resource consumption. Based on this, a improved Residual Cumulative Resampling (RCR) method based on stratified rough sorting is proposed in this paper. The stratified rough sorting method adds the operation of stratifying the particles according to the weight on the basis of the original rough sorting method to ensure that the large weight particle set always lies above the small weight particle set in the sorting process, which improves the sorting stability and ensures the tracking accuracy. The experimental results show that the RCR method based on stratified rough sorting outperforms the original rough sorting method in both OSPA distance and target number estimation, and the specific effect is comparable to Multinomial Resampling and Systematic Resampling, which can improve the tracking accuracy while preserving the simple structure of rough sorting.