Research On Unmanned Surface Vehicle Path Planning Based On Improved Locking Sweeping Method

With the proposal of the national marine development strategy,the research and development of unmanned surface vehicle(USV)has become the focus of more and more scholars.It has the characteristics of high speed,intelligence and unmanned.It has wide application and important research value,and path planning technology is the prerequisite for completing various tasks.This paper taking USV as the main research direction,starting from the aspects of USV large-scale autonomous navigation,dynamic obstacle avoidance and realtime replanning,focuses on the global path planning of USV in static environment and the local path planning of USV in dynamic environment.The specific contents are as follows:Firstly,the locking sweeping method(LSM)has obvious advantages in solving the path planning problem of USV in large-scale and complex environment,but it also has problems of local optimization and low computational efficiency in the treatment of obstacles in local path planning.To solve these problems,this paper improves the LSM algorithm by introducing velocity obstacle method(VO)and proposes LSM-VO algorithm.LSM-VO algorithm can make full use of the advantages of LSM algorithm,such as relatively optimal global planning path and rapid replanning after obstacle avoidance,avoid the disadvantages of low efficiency of dynamic obstacle avoidance reconstruction potential field,and make full use of the good realtime performance of dynamic obstacle avoidance of VO method.Through qualitative and quantitative comparative experimental analysis,this method has the advantages of high realtime and smooth planned path,and meets the needs of USV path tracking and temporary obstacle avoidance decision under real navigation sea conditions.Secondly,aiming at the problem of multi dynamic obstacle avoidance in the Bay scene with heavy traffic and in the process of entering and leaving the port,the multi dynamic obstacle avoidance strategy based on VO algorithm is studied and applied to LSM＿VO algorithm deals with the avoidance problem of multiple moving ships.Through the comparative analysis of simulation experimental results,the effectiveness and feasibility of the algorithm are verified.It can safely complete the whole task process,quickly complete the safe heading selection in the situation of multiple dynamic obstacles,and carry out online replanning in combination with the global guidance potential field after avoidance,which effectively improves the reliability and calculation efficiency of the algorithm.It can solve the problems of large-scale bow turning and zigzag route planning in dynamic obstacle avoidance of LSM algorithm,and meet the requirements of real-time dynamic obstacle avoidance of USV autonomous navigation and the realization of real-time ship maneuver.Finally,combined with the experimental data of real USV maneuverability,the influence of course angle of USV,LSM＿VO algorithm is further improved and LSM-CVO algorithm is designed.Combined with the maneuverability of real USV,multi motion obstacle simulation tests at different speeds are carried out.The high-density obstacle area is designed for global planning and local collision avoidance simulation environment and multi obstacle collision avoidance in Bay scene.The simulation results are compared and analyzed to verify that the improved algorithm meets the requirements of real USV maneuverability.