Hybrid Differential Evolution Particle Filter for Nonlinear Filtering

作者：Zhang, Chaozhu^*; Li, Lin

来源：Applied Computational Electromagnetics Society Journal, 2014, 29(12): 1133-1139.

• Hybrid differential evolution
• nonlinear filtering
• particle filter
• radar tracking
• simulated annealing algorithm

In this paper we propose a novel method for solving the nonlinear problem of the radar target tracking. The algorithm consists of a Particle Filter (PF) which employs the Unscented Kalman Filter (UKF) to generate the importance proposal distribution, and adopts the Hybrid Differential Evolution (UDE) algorithm based on Simulated Annealing (SA) algorithm as the resampling scheme. Firstly, the Importance Distribution (ID) which contains the newest measurements is constructed by the UKF. In addition, the UKF generates proposal distributions that match the true posterior more closely. Secondly, to solve the particle degeneracy and impoverishment phenomenon, the sampling particles are resampled by the HDE algorithm. The mutation and crossover steps of the Differential Evolution (DE) algorithm are executed to generate the trial vectors. Then the selection step is replaced by the Metropolis criterion of the SA algorithm. The proposed algorithm combines the advantages of the SA algorithm with the DE algorithm. It not only has superior estimation performance, but also the convergence speed is fast. Simulation results demonstrate that the proposed algorithm outperforms the standard PF, the Auxiliary Particle Filter (APF), the Regularized Particle Filter (RPF) and the Particle Filter based on Differential Evolution (PFDE).

• 出版日期2014-12
• 单位哈尔滨工程大学