An Active Fault Detection for Unmanned Surface Vehicles With Minor Fault

This article proposes an active fault detection method for Unmanned Surface Vehicles (USVs) with uncertain bounded disturbance to achieve the minor faults detection of USV. In practice, minor faults are the early cases of normal faults with the characteristic of low amplitude and difficulty to detection. At first, a set-membership estimation approach is used to describe the area topology of the nominal model of USVs and the fault models. Then, an auxiliary input signal is designed to enhance the character of minor faults, and the minor faults can be separated from the considered USVs in spite of uncertain bounded disturbances. Next, the considered problem is transformed into a nonconvex optimization problem, and the optimal auxiliary signal is obtained via solving the mixed integer quadratic programming (MIQP). Finally, a case study of the USVs is used to show the effectiveness of the proposed method.