A hierarchical multiple-model approach for detection and isolation of robotic actuator faults

Modern robotic systems perform elaborate tasks in a complicated environment and have close interactions with humans. Therefore fault detection and isolation (FDI) schemes must be carefully designed and implemented on robotic systems in order to guarantee safe and reliable operations. In this paper, we propose a hierarchical multiple-model FDI (HMM-FDI) scheme for robotic actuator faults. The proposed algorithm performs FDI in stages and refines the associated model set at each stage. Consequently only a small number of models are required for a wide class of faults, including abrupt faults, incipient faults, and simultaneous faults. Experiments are conducted to show that the HMM-FDI scheme can successfully and immediately detect and isolate various types of actuator faults.