基于Sigma点H_∞滤波的拟蒙特卡罗粒子滤波算法

Quasi-Monte Carlo particle filter algorithm based on sigma point H_∞

在滤波算法中,用Sigma点H∞滤波来产生重要性概率密度函数,由于Sigma点H∞滤波对不确定观测噪声具有较强的鲁棒性,而且在滤波过程中考虑了最新的观测值,因此由其产生的重要性函数更逼近于真实的后验概率分布。同时在重采样阶段,利用拟蒙特卡罗重采样算法进行重采样,有效地克服了粒子退化现象并提高了状态估计精度。仿真结果表明了所提算法的可行性和有效性。

A new filtering algorithm is proposed.In this algorithm,the probability density function is generated by the sigma point H∞filter.The sigma point H∞filter has very high accuracy and strong robustness to uncertain observation noise,and the filter algorithm integrates the new observation.So the generated probability density function can reasonably approximate the real posterior probability distribution of the system state.At the resampling step,the degeneration problem is effectively overcome and the accuracy of state estimation is improved by using the quasi-Monte Carlo-based resampling algorithm.Simulation results demonstrated the feasibility and effectiveness of the proposed algorithm.