Safety-Critical Disturbance Rejection Control for Capsizing Prevention of an Unmanned Sailboat

The capsizing of unmanned sailboats primarily results from strong winds, high waves, and improper maneuvering. However, most existing unmanned sailboat control approaches do not explicitly address roll dynamics or capsizing prevention. This paper investigates the capsizing prevention of an unmanned sailboat in the presence of the model uncertainties and unknown environmental disturbances. A safety-critical disturbance rejection control method is proposed to ensure the safe maneuvering of unmanned sailboat. Specifically, an extended state observer is presented to not only filter the velocity measurements, but also estimate the total disturbance composed by the model uncertainties and unknown environmental disturbances. By using the backstepping technology, the virtual guidance and control law is then derived. Both the control barrier function (CBF) and the input-to-state safe CBF are introduced to construct the safety condition. Based on the safety constraint, a quadratic programming problem is formulated to obtain an optimal safety-critical control law. Furthermore, the input-to-state stability of the closed-loop system is proven by Lyapunov theory, while its input-to-state safety is guaranteed. Comparative simulations under different disturbance scenarios using the developed high-fidelity unmanned sailboat model demonstrate the effectiveness of the proposed control method in capsizing prevention.