Research On Decision-Making Method Of Collision Avoidance Behavior For Unmanned Underwater Vehicle In Unknown Environment

In the unknown environment,the ability of autonomous avoidance of the unmanned underwater vehicle(UUV)is the premise and foundation to ensure its safety to complete its mission.Usually,based on the environmental information detected by the sonar,UUV can make the decision by itself,and then sails to the target without collision.However,because of the uncertainty and time variability of environment,the complexity of autonomous avoidance behavior is increased and it also becomes more difficult for UUV to avoid the collision in-timely and effectively.It is of great theoretical and practical significance to study the decision-making method of UUV's autonomous avoidance behavior in unknown environment.The main work of this paper is as follows:Firstly,based on the research status of UUV's autonomous avoidance behavior,this paper starts out with the UUV's collision avoidance problem in the unknown static environment,the environmental model is established and a algorithm based on UUV reactive avoidance behavior decision is proposed and designed.Secondly,according to the working principle of the multi-beam sonar,a polar coordinate system is established to judge the environmental information quickly.According to the actual demand,the minimum turning radius is added into the autonomous avoidance method of the UUV,and the judgments and analysis for the feasibility of UUV's predicted path was made.Then,for the problem of collision avoidance in unknown static obstacles,a reactive behaviors collision avoidance algorithm is applied.The sub-target points in the path are selected by evaluation function.For the problems of minimal value in the algorithm,which is also the characteristic of most real-time collision avoidance algorithms,a boundary memory method was proposed.Then,the feasibility of UUV avoiding unknown static collision is proved by simulation experiments.Finally,based on the reactive algorithm,a collision avoidance strategy is designed for unknown dynamic obstacles.Though the unknown dynamic and static obstacles were added in the simulation experiment,the results were still very good.Experiments show that the collision avoidance algorithm proposed in this paper is effective and feasible.