Adaptive Observer-Based Fault Reconstruction Method for Unmanned Surface Vehicles

This paper studies the fault estimation problem for an unmanned surface vehicle (USV) with simultaneous sensor and actuator faults. Through introducing a linear transformation and a virtual measurable variable, two adaptive observers are implemented to solve this issue, where the first one can generate the exact estimations of the actuator fault together with the disturbances, and the second one can reconstruct the sensor fault. The proposed method reduces the design complexity compared with the augmented descriptor observer, especially the sensor fault can be transferred into the form of actuator fault. Then, the stable condition of the resulting error systems is also presented. Finally, the numerical simulation is given to validate the fault estimation method of the USV.