噪声统计特性未知时的鲁棒卡尔曼滤波算法设计

摘要

This paper is concerned with the problem of a robust Kalman filter(RKF) design when noise covariance is unknown in stochastic linear systems. A novel design criterion for RKFs is proposed, and its rationality is analyzed. Based on the criterion, the design of a RKF is transformed to solving a linear matrix inequality(LMI). The results are validated by simulations.%本文研究了噪声统计特性未知时的鲁棒卡尔曼滤波算法(RKF)设计问题.首先,提出了一种新的RKF算法设计条件,并分析了其合理性;其次,从RKF算法设计条件出发研究了RKF算法的设计问题,把RKF算法的设计过程转化为计算一组线性矩阵不等式(LMI)的可行解;再次,研究了LMI可行解的计算问题,并通过计算该LMI的可行解设计了一种RKF算法;最后,通过仿真验证了所设计的RKF算法的有效性.