Automatic Ship Collision Avoidance Algorithm Based on AIS Data Processing and Improved DWA Algorithm

Abstract

This study proposes a path planning algorithm for maritime navigation that combines global trajectory extraction from a ship’s automatic identification system (AIS) data with local dynamic collision avoidance using the dynamic window approach (DWA). AIS data from the target region is clustered to identify habitual navigation points, which are connected via Dijkstra’s algorithm to form an undirected trajectory between the Start and the Goal. The trajectory is further refined using a dynamic programming (DP) algorithm. An improved DWA algorithm is developed to enhance local collision avoidance, incorporating wind and current factors into the ship motion model. A route deviation evaluation function is introduced to facilitate route recovery after obstacle avoidance and minimize path length. Additionally, a collision avoidance rule evaluation function is integrated to address diverse maritime encounter scenarios, enabling the planning of safe and regulation-compliant paths that better align with real-world navigation requirements, further minimizing potential risks. The proposed method ensures efficient integration of global and local path planning, enhancing both safety and navigational efficiency. Simulation experiments validate the effectiveness of the algorithm, demonstrating its potential for application in maritime navigation systems.