应用粒子群优化的高斯粒子滤波
Gaussian particle filtering using particle swarm optimization

针对高斯粒子滤波（ＧＰＦ）在多峰高斯假设条件下不能满足贝叶斯估计精度的问题，提出一种基于粒子群优化的高斯粒子滤波算法（ＰＳＯ－ＧＰＦ）． 该算法用粒子群优化算法更新高斯建议分布的参数，解决粒子退化和多峰高斯下的粒子精度问题． 同时，带压缩因子的粒子群优化算法能有效平衡粒子的全局探测与局部开采． 实验结果表明，新算法的滤波精度比高斯粒子滤波精度平均可提高93.9%，具有更高的稳定性．
Under the condition of multi-peak Gaussian hypothesis，general Gaussian particle filtering (GPF) dissatisfies the accuracy of Bayesian estimation. This paper proposes a Gaussian particle filtering algorithm using particle swarm optimization (PSO- GPF). The particle swarm optimization (PSO) algorithm is applied to update the parameters of the Gaussian proposal distribution to ease the particle degradation and solve the precision in the multi-peak Gaussian case. Meanwhile，the PSO algorithm joined the compression factors effectively balances the global detection and local mining of the particles. The experimental results show that the precision of the new filtering improves averagely by 93.9% and has higher stability when compared with the GPF