The Path Planning Method For Unmanned Surface Vehicles In Compliance With COLREGS

In recent years,unmanned driving technology has developed rapidly.Due to various complex factors,unmanned surface vessels(USV)have lagged behind in development,but it is foreseeable that they will play a crucial role in the field of navigation.As a key technological node for USV,a well-developed path planning algorithm has significant research significance.The existing path planning algorithms for USV have significant drawbacks: manned vessels navigating on the sea follow the unified technical specifications for collision avoidance behavior as stipulated in the "International Regulations for Preventing Collisions at Sea(COLREGS)." However,the path planning methods for USV have not adapted to these regulations,resulting in different USV adopting different avoidance behaviors when encountering dynamic obstacles.As USV cannot currently replace manned vessels,these different avoidance behaviors pose significant safety risks.To address these issues,this paper improves the fast marching method to ensure that the planned path complies with the COLREGS allowing unmanned surface vessels to adopt the same collision avoidance behavior as USV and increasing the safety of maritime navigation.This paper first introduces the traditional fast marching method,which constructs a potential field and then uses gradient descent to find the shortest path.This method has the potential to solve the problem of path planning algorithms not being able to adapt to the COLREGS.However,the paper also discusses the drawbacks of applying the fast marching method to path planning for USV,which include paths being too close to obstacles and having too many turning points.Therefore,the paper introduces obstacle time field functions and gradient turning point evaluation functions based on the fast marching method.The obstacle time field function creates a time field around obstacles to keep the planned path away from them and increase safety.For dynamic obstacles,the planned path is constrained by the offset obstacle time field,allowing it to avoid dynamic obstacles and comply with the COLREGS.The gradient turning point evaluation function replaces gradient descent,reducing the number of turning points in the planned path while still allowing it to reach the destination quickly.Finally,experiments are designed to solve the important parameters in the obstacle time field function and gradient turning point evaluation function,and a reasonable method for solving these important parameters is obtained.The improved algorithm was designed and simulated on the Matlab platform,and compared with the traditional fast marching algorithm.The results showed that the improved fast marching algorithm,which uses obstacle time-field and gradient turning point evaluation functions,can increase the safety and efficiency of USV while not sacrificing the complexity of the planned path.Simulation experiments were also designed to verify the ability of the algorithm to avoid dynamic obstacles in compliance with the COLREGS.The experiments confirmed that the offset of the obstacle time-field can enable the planned path to avoid dynamic obstacles in accordance with the regulations.