一种基于粒子群优化的粒子滤波改进算法

粒子群优化粒子滤波算法能有效改善粒子退化问题,但其适应度函数受量测噪声方差影响较大,限制了滤波精度的提高.为此,提出了一种基于粒子群优化的粒子滤波改进算法.该算法给出一种新的适应度函数,用当前状态估计值与各粒子状态的差值大小作为评价标准,使得最终优化粒子受噪声方差影响减小,在量测模型精度高的场合中提高了滤波精度.理论分析及仿真结果表明,本文所提算法的滤波性能优于标准粒子滤波与粒子群优化粒子滤波算法.

Improved Algorithm of Particle Filtering Based on PSO

Although the particle filtering algorithm of particle swarm optimization （PSO） can effectively lower the particle degeneracy, the fitness function is affected greatly by measured noise variance, which bounds the improvement of filtering precision. Therefore, an improved algorithm for particle filtering based on particle swarm optimization （PSO-PF） is proposed in this paper. As for this algorithm, a new fitness function is put forward, and the size of difference between the current state estimation value and the state of each particle is taken as the evaluation standard, thus, allowing finally the optimized particles is less affected by noise invariance and increasing the filtering precision on the occasion of measured model with higher precision. Theoretical analysis and simulation findings show that the filtering performance of this proposed algorithm is superior to those of standard particle filtering and particle swarm optimization algorthms.