Research On Cooperative Path Planning And Dynamic Obstacle Avoidance For Multiple USVs

Compared with the popularity of unmanned aerial vehicles,the application of unmanned ships is still in its infancy and is becoming a research hotspot nowadays.The path and obstacle avoidance trajectory planning algorithm is a core part of realizing the safe navigation of unmanned ships in the complex ocean environment.In the case of multi-vessel coordination,it is also necessary to consider the constraints of navigation time and voyage.This article is mainly to deal with these issues,and the main content is divided into three parts:global path planning,single unmanned ship autonomous obstacle avoidance trajectory planning,and multi-vessel coordinated trajectory planning.In this thesis,the improved Ant Colony Optimization(ACO)is used as the basic algorithm for the global path planning of the unmanned surface vehicle.This algorithm combines the ant colony system(ACS)method and adaptive strategy to adjust the its convergence accuracy and speed.In the autonomous obstacle avoidance trajectory planning module of the unmanned ship,a collision risk evaluation function based on the minimum encounter distance and the shortest encounter time is designed,and the requirements of the existing maritime regulations in collision avoidance are analyzed and sorted out.According to the principle of speed obstacle,the relative speed direction of the unmanned ship and the obstacle to achieve obstacle avoidance is converted into the desired unmanned ship heading,and the solution space of the unmanned ship is further restricted in combination with the international maritime collision avoidance rules.Finally,the dynamic window approach is improved to realize that the unmanned ship can avoid obstacles at a fixed speed and return to the original global path after resuming under the constraints of collision avoidance rules at sea.In the case of multi-ship cooperation,this thesis chooses the differential evolution algorithm to deal with the task assignment problem.Taking the triplet consisting of the corresponding number of the unmanned ship and the task performed and the voyage cost between them as a gene locus,a mutual mapping strategy between the solution space and the cost space is designed,and the differential evolution operation is performed through the cost space.The dynamic mutation rate and hybrid difference strategy are used to balance the global exploration and local development capabilities of the algorithm,and the scaling factor is improved to further improve the convergence speed and developability of the algorithm.Finally,based on the constraints and optimization objectives,the effectiveness of the improved differential evolution algorithm in dealing with the multi-USV cooperative task assignment problem is verified.Finally,this paper combines global path planning and single-USV obstacle avoidance trajectory planning modules to simulate and verify the effectiveness of multiple USVs in cooperative execution of tasks and autonomous collision avoidance.