多率系统Kalman滤波算法的鲁棒性分析

Robustness Analysis of Kalman Filtering Algorithm for Multirate Systems

多率系统Kalman滤波算法是多率采样系统中对多源观测进行融合的重要手段.基于化工过程的多率采样特点,给出了多率Kalman滤波算法,分析了该算法在模型失配情况下的鲁棒性.在给定的假设条件下,通过对滤波误差变化规律的分析,给出了多率Kalman滤波稳定与发散的基本判定方法.针对一类典型系统,推导出了滤波稳定与发散判据.通过仿真对该判据进行了验证,仿真结果表明所提出的滤波鲁棒性分析方法可以用于算法的实际应用.

The multirate Kalman filtering algorithm is an important method of fusing multi-source observations in multirate systems. The multirate Kalman filtering algorithm is given based on the multirate-sampling character of chemical processes, and the robustness of the algorithm under model mismatches is investigated. By analyzing the evolution pattern of the filtering error under given hypothesis, a basic judging method of multirate Kalman filtering divergence is proposed. Then, a criterion of the filter divergence property for a class of typical systems is derived. Simulation result shows that the proposed method can be applied to the multirate Kalman filtering algorithm.