Actuator fault diagnosis for multi-sensor systems with asynchronous IMM and MLE

This paper proposes a new actuator fault detection and estimation method for dynamic system with multiple sensors based on asynchronous Interactive Multiple Model (IMM) and Maximum Likelihood Estimation (MLE). Both total and partial actuator faults are considered. Asynchronous IMM filtering is performed to the model set composed of the normal system model and all fault models corresponding to possible fault situations. Model probabilities as indicators of occurrences of various fault situations are obtained and utilized to detect and locate the fault. After that, the fault factor is estimated based on the MLE algorithm and the analytic solution is given out. Numerical simulations are presented to illustrate that the proposed approach can not only detect and isolate the fault correctly, but also have improved estimation performance of fault factor.