Sensor Fault Identification and Decentralized Fault-Tolerant Control of Reconfigurable Manipulator Based on Sliding Mode Observer

An approach of fault-tolerant control for sensor fault of reconfigurable manipulator based on sliding mode observer and decentralized adaptive fuzzy is proposed. The dimension of original system is extended by introducing a first order filter such that sensor fault is equated to actuator fault. The sliding mode observer is designed for the new system which coupling interconnection terms of the joints and modeling uncertainties are estimated by fuzzy system. Then sensor fault can be reconstructed by using the state estimations and system output to implement fault-tolerant control. Simulation results of two different configurations of 3-DOF reconfigurable manipulator verify the validity of the proposed approach.