基于混沌的改进粒子群优化粒子滤波算法

针对基本粒子滤波(PF)算法存在的粒子退化和重采样引起的粒子多样性丧失,导致粒子样本无法精确表示状态概率密度函数真实分布,提出了一种基于混沌的改进粒子群优化(PSO)粒子滤波算法。通过引入混沌序列产生一组混沌变量,将产生的变量映射到优化变量的区间提高粒子质量,并利用混沌扰动克服粒子群优化局部最优问题。利用单变量非静态增长模型(UNGM)在高斯噪声和非高斯噪声环境下将该算法与基本粒子滤波和粒子群优化粒子滤波(PSO-PF)的性能进行仿真比较。结果表明:该算法的性能在有效粒子数和均方根误差(RMSE)等参数都优于基本粒子滤波和粒子群优化粒子滤波,改善了算法的精度和跟踪性能。 

Improved particle filter algorithm based on chaos particle swarm optimization

To solve the degeneracy phenomenon and the sample impoverishment problem of basic particle filter (PF) algorithm, which makes the particles of PF algorithm unable to express the real distribution of probability density function, a novel PF algorithm based on chaos particle swarm was proposed. Chaos sequence was adopted in this proposed algorithm. The chaos sequence was used to generate a set of chaotic variables, which was mapped to the interval of optimization variables to improve the quality of particles. And chaos perturbation was utilized to overcome the search being trapped in local optimum for particle swarm optimization (PSO) algorithm. The univariate nonstationary growth model (UNGM) was used for simulation to compare the proposed algorithm with basic PF and particle swarm optimization particle filter (PSO-PF). Under the conditions of Gaussian and non-Gaussian noise, the performances of the proposed algorithm had been verified by the simulation. The results show that the performances of the number of effective particles and root mean square error (RMSE) in the algorithm are better than the performances of the basic PF and the PSO-PF algorithm. Therefore, the accuracy and tracking performance of PF are improved.