A Tutorial on Particle Filters for Fault Detection and Diagnosis of Hybrid Systems

Hybrid dynamic systems have recently been a hot research topic. As the complexity of such systems is increasing day-to-day and system sizes are continually expanding, it can cause significant financial loss and loss of life in the event of system collapses. Therefore, ensuring the reliability and security of systems is of major importance in societal and economic terms. The research of particle filtering has made tremendous advances in the last three decades and has been demonstrated to be a good way to estimate the states of systems. So the combination of particle filtering with hybrid systems can be a promising method in fault detection and diagnosis. In this work, particle filtering algorithms are introduced and the emphasis is in their usage in hybrid systems fault detection and diagnosis. The review reveals the strengths and weaknesses of these particle filtering methods, and some examples are given to illustrate the practical application in detection and diagnosis based on these methods. Finally, future research directions and the difficulties that remained to be resolved are examined. It is anticipated that this paper will help researchers in this field understand the methodology better and provide a new concept and method in hybrid systems fault detection and diagnosis.