A neural network based actuator fault detection and diagnostic scheme for a SCARA manipulator

Abstract

One of the most critical components of a robotic system is the actuator, which undergoes a lot of wear and tear and may lead to its failure. In order to monitor such a system, we propose a neural network-based fault detection and diagnosis scheme for actuator failures in robotic manipulators. A single detection and diagnostic observer is utilized for online failure assessment and the weights of the failure online approximators are adaptively updated using Lyapunov re-design methods. The fault detection scheme is implemented for a SCARA manipulator and simulation results are presented.