Development of Fault Detection and Identification System for Thrusters of Underwater Robots

Abstract

The paper is dedicated to solving of on-line fault detection and identification for thrusters of underwater robots. To detect and estimate the possible typical faults, bank of sliding mode observers is proposed. The new approach to construct such a bank of observers, based on the reduced model of the initial systems, is suggested and developed. This approach allows reducing complexity of the observers. The effectiveness of the proposed system was confirmed by computer simulation.