A Novel Adaptive Resampling Particle Filtering Algorithm

Resampling is employed to alleviate the phenomenon of particle degradation in a particle filter. However, resampling weakens the diversity of particles, and results in inadequate convergence accuracy. To coordinate the conflict between the validity and diversity of the particle sets, an adaptive multi-level resampling based particle filtering algorithm (ARPF) is proposed in this paper. The particle resampling weight space algorithm is divided into multiple levels, and the diversity of the resampled particles is measured by the rank entropy to conduct resampling. The experimental results show that the multi-level resampling particle filtering algorithm can improve the accuracy, and is suitable for high precision system filtering.