Adaptive robust extended Kalman filter for nonlinear stochastic systems

作者：Xiong K^*; Zhang H; Liu L

来源：IET Control Theory and Applications, 2008, 2(3): 239-250.

DOI：10.1049/iet-cta:20070096

The authors analyse the error behaviour of the robust extended Kalman filter (REKF) for nonlinear stochastic systems. On the basis of some standard results about the boundedness of stochastic processes, it is specified that stability of the REKF cannot be guaranteed. In order to solve this problem, a novel method is proposed to design the REKF so that the sufficient conditions to ensure filter stability will be fulfilled. Furthermore, an adaptive scheme is adopted to automatically tune the error covariance matrix in response to the changing environment. Numerical example shows the superiority of the proposed adaptive REKF over the usual extended Kalman filter (EKF), the REKF and the adaptive EKF.

• 出版日期2008-3
• 单位北京控制工程研究所; 北京航空航天大学