A Hybrid Multi-Target Path Planning Algorithm for Unmanned Cruise Ship in an Unknown Obstacle Environment

Area of Science:

• Robotics
• Artificial Intelligence
• Marine Engineering

Background:

• Multi-target path planning for unmanned surface vehicles (USVs) in unknown environments presents significant challenges.
• Efficiently determining optimal routes and navigating complex, obstacle-filled waters is crucial for autonomous marine operations.

Purpose of the Study:

• To develop a novel hybrid algorithm for efficient multi-target path planning for unmanned cruise ships in unknown lake environments.
• To improve the speed, efficiency, and path smoothness of autonomous navigation systems.

Main Methods:

• The study proposes a two-part hybrid algorithm: 1) Transforming the multi-target problem into a Traveling Salesman Problem solved by an improved Grey Wolf Optimization (GWO) algorithm with a Beta function for faster convergence. 2) Employing an improved D* Lite algorithm with an enhanced heuristic function for efficient path planning between targets, reducing node expansion and smoothing paths.
• The improved GWO algorithm enhances convergence speed by optimizing the convergence factor using the Beta function.
• The improved D* Lite algorithm enhances search speed and path smoothness by reducing expanded nodes via heuristic function modification.

Main Results:

• Experimental validation demonstrated the proposed hybrid algorithm's strong applicability and high effectiveness.
• Comparative analysis against four other algorithms in ordinary and complex environments confirmed the superiority of the proposed method.
• The algorithm successfully planned efficient and smooth paths for unmanned cruise ships in unknown obstacle-laden environments.

Conclusions:

• The hybrid multi-target path planning algorithm effectively addresses the challenges of autonomous navigation in unknown waters.
• The integration of improved GWO and D* Lite algorithms offers a significant advancement in USV path planning capabilities.
• The proposed method provides a robust and efficient solution for real-world applications of unmanned marine vehicles.