University of Hertfordshire

Collision avoidance and risk assessment are open problems to be practically addressed in maritime transportation. In high-speed vessels this problem becomes more challenging due to manoeuvring and reaction time constraints. Here, a reactive collision avoidance and risk assessment technique with fuzzy weighting functions are proposed for a relatively high-speed autonomous catamaran. To follow paths between predefined waypoints, a Line of Sight (LOS) technique with Cross Tracking Error (CTE) is utilised. Besides, a new collision risk index is introduced based on fuzzy weighting functions. To perform formal maritime decision making, the standard marine COLlision REGulations (COLREGs) are incorporated into the algorithm. Furthermore, a simplified Closest Point of Approach (CPA) formulation is presented. The proposed framework is simulated on a realistic model of a vessel including input and non-holonomic constraints and disturbances. Simulation results for various encounter scenarios demonstrate the merits of the proposed method.