动态噪声特性未知系统的多模型自适应卡尔曼滤波

Multiple model adaptive Kalman filter for the system without the knowledge of process noise

摘要: 针对常规自适应卡尔曼滤波器存在过渡过程差的问题,基于一个给定的指标切换函数,采用多个基于不同动态噪声协方差矩阵的卡尔曼滤波器和一个常规自适应卡尔曼滤波器共同组成多模型自适应卡尔曼滤波器.与常规自适应卡尔曼滤波器相比,多模型自适应卡尔曼滤波器可以在保持原有自适应滤波器性能的基础上极大地改善瞬态响应.

Abstract: To solve the bad transient response of a conventional adaptive Kalman filter, multiple Kalman filters based on different noise covariances and a conventional adaptive Kalman filter (AKF) were used to form a multiple model adaptive Kalman filter (MMAKF) by using a switching index function. Compared with a conventional AKF, the MMAKF could improve the transient response greatly without losing the characteristic of the conventional AKF.