MEMS gyroscope fault detection and elimination for an underwater robot using the combination of smooth switching and dynamic redundancy method

Abstract

Micro Electro Mechanical System (MEMS) gyroscope sensors are used in different underwater vehicles to determine their attitude as the input of the controller system to adjust the attitude to the desired value. Fault detection and elimination are necessary for MEMS gyroscope sensors to avoid disruption of the system performance. In this paper, it is tried to study different types of gyroscope sensor faults which can be detected and eliminated by designing a new proposed dynamic redundancy scheme. The proposed method is consisted of two MEMS gyroscope sensors and is modified by considering underwater robot dynamics in threshold tuning via Monte Carlo simulations. The switch between the primary and secondary gyroscope sensors is designed by smoothing methods to eliminate undesirable behavior of switching kick. Finally, numerical and experimental simulations are performed by a three degree of freedom table to show that the proposed method is efficient for the underwater robot exposed to gyroscope sensor faults.