Self-recovering extended Kalman filtering algorithm based on model-based diagnosis and resetting using an assisting FIR filter

Abstract

This paper proposes a new intelligent filtering algorithm called the self-recovering extended Kalman filter (SREKF). In the SREKF algorithm, the EKF׳s failure or abnormal operation is automatically diagnosed using an intelligence algorithm for model-based diagnosis. When the failure is diagnosed, an assisting filter, a nonlinear finite impulse response (FIR) filter, is operated. Using the output of the nonlinear FIR filter, the EKF is reset and rebooted. In this way, the SREKF can self-recover from failures. The effectiveness and performance of the proposed SREKF are demonstrated through two applications – the frequency estimation and the indoor human localization.