Robust fault detection of robotic systems: new results and experiments

Abstract

This paper proposes a new approach, called robust nonlinear analytic redundancy (RNLAR) technique, to sensor and actuator fault detection for input-affine nonlinear multivariable dynamic systems that include most robotic systems. In this approach, both model-plant-mismatch (MPM) and process disturbance are considered during fault detection. The proposed RNLAR can be used to design primary residual vectors (PRV) for nonlinear systems to detect faults. A nonlinear PRV design method to detect sensor and actuator faults is proposed where the PRVs are made highly sensitive to the faults and less sensitive to the MPM and process disturbance. Experimental results on a PUMA 560 manipulator are presented to justify the effectiveness of the proposed RNLAR technique