混合粒子流滤波的非线性系统参数估计算法

A Nonlinear System Parameter's Estimation Algorithm Based on Hybrid Particle Flow Filter

为了提高强环境噪声下非线性系统估计性能,基于粒子流滤波对非线性系统估计能力强的特点,文中首先利用粒子流滤波粗估计状态向量;然后,利用卡尔曼滤波平滑由强环境噪声所导致的状态向量估计误差;最后,得到混合粒子流滤波算法。对转移方程为线性而测量方程为非线性的系统估计仿真实验表明:文中算法的参数估计精度高于普通粒子流滤波算法和粒子滤波算法,计算复杂度和普通粒子流滤波算法相当且低于粒子滤波算法。

In order to improve estimation performance of the nonlinear system under strong environmental noise, the state vector is roughly estimated by particle flow filter firstly since it is good for handling nonlinear system estimation problem. Then the state vector＇s estimation error, which is caused by the strong environment noise, is smoothed by a Kalman filter. Finally the hybrid particle flow filter is gotten. The results of simulation for the system estimation consisting of linear transfer equation and nonlinear measurement equation show that the estimation accuracy of the proposed algorithm is higher than that of the standard particle flow filter and the particle filter, computational complexity of proposed algorithm is the same as standard particle flow filter and is lower than that of the particle filter.