Modeling and reactive navigation of an autonomous sailboat

Among Autonomous Surface Vehicles (ASV), sailboat robots could be an efficient solution for long term missions and semi-persistent presence in the oceans since they rely on renewable energy like solar and wind energies. However, steering of such vehicles presents an inherent difficulty since the thrust force heavily rely on the wind (speed and direction) and sail angle as well as on the underlying complex dynamic properties of the ship, mainly due to the aero and hydrodynamic forces. This paper proposes a new reactive path planning method which takes into account the so called no-go zones and turns them into virtual obstacles. A potential field algorithm is then applied to drive the sailboat towards the goal while avoiding obstacles even when the goal is located directly downwind or upwind. Simulations, using a boat specific dynamic simulator, show that the proposed method can successfully drive an autonomous boat between a set of predefined waypoints under time varying wind conditions. This method is computationally efficient and is suitable for on-board real-time implementation.