Active fault diagnosis for AUV over-actuated systems based on UIOs and control allocation

The navigation performance of autonomous underwater vehicles (AUVs) relies mainly on the propulsion system, which makes the study of the safety and reliability of the propulsion system highly significant. This paper investigates an over-actuated AUV propulsion system with six degrees of freedom and presents an active fault diagnosis framework that combines unknown input observers (UIO) with control allocation. The simulation results show that the proposed integrated methodology successfully performs fault detection, magnitude estimation, and isolation for AUV over-actuated systems. In particular, since no auxiliary signals are added, it is possible to ensure normal operation of the system by only internally adjusting the allocation of the original inputs.