Actuator Fault Tolerant Control for a Remotely Operated Vehicle based on Adaptive Extended Kalman Filter

In this paper, we propose a strategy for detecting and isolating actuator faults in an overactuated Remotely Operated Vehicle (ROV) with six degrees of freedom. Fault detection and isolation are based on the Adaptive Extended Kalman Filter (AEKF), which is a recent extension of the well-known Kalman Filter designed for nonlinear dynamics and additive disturbances estimation. The residuals generated by the AEKF act as directional residuals, and fault isolation is performed by calculating the cosine similarity between the residuals and the columns of the control effectiveness matrix. The decision is then used in a simple fault-tolerant control allocation algorithm, without the need to alter the control law. Simulation results show the effectiveness of the estimation method in the presence of thruster failures.