Research On Intelligent Evasion Method Of USV Responding To Active Threat

Unmanned Surface Vehicle(USV)has become a research hotspot following the unmanned aerial vehicles(UAV)because of its high mobility,strong concealment and low cost.With the development of sensors,computers and other technologies,USV can be applied to complex tasks such as coastal patrols,anti-submarine tracking,surveillance,and so on in recent years.The autonomous ability to defend itself from the threat is the basis for the USV to complete the task of confrontation,which also reflects its intelligence level.There may be some problem about collision,interception that may be encountered in the process of patrolling,monitoring and reconnaissance in the disputative sea area.For this problem,the paper builds a two-layer model of online real-time trajectory planning and reactive obstacle avoidance,which can guide the USV to deal with different threats.Firstly,comprehensively analyzing the threat avoidance algorithm,collision avoidance algorithm and penetration strategy of domestic and international unmanned vehicles and based on the existing collision avoidance system,this paper puts forward a hierarchical threat avoidance strategy.Through the study of the shortest meeting time and the shortest distance between the two ships,the method of calculating the risk of collision and the method of determining the type of danger are put forward.And the selection and conversion of the reference coordinate system needed for the USV trajectory planning and collision avoidance are analyzed.In order to achieve the online unmanned trajectory planning and collision avoidance of static and low speed obstacles,real-time Rapidly Exploring Random Tree(RRT)algorithm is selected as the first layer of unmanned trajectory planning and collision avoidance of USV.The shortcomings of RRT algorithm are obvious: randomness,no objectivity,poor smoothness of the generated path,and no guarantee of optimality or asymptotic optimization.In order to solve the shortcomings of RRT algorithm,this paper proposes the target bias strategy,which can increase the maneuverability of USV,introduce the path evaluation function and optimize node connection strategy,thus the improved RRT algorithm can search for the optimal trajectory quickly to avoid the threat under the premise of satisfying the maneuverability of the USV.Finally,the ordered waypoint generated by the improved RRT algorithm is fitted by Catmull-Rom interpolation.In order to cope with situations where the USV may occur during the execution of a mission,such as naval confrontation,interception,impact and so on,and these are high dynamic proactive threats,two algorithms are combined as the avoidance algorithm under the USV collision urgency,one algorithm is a dangerous avoidance algorithm Velocity Obstacles algorithm(VO)in the first order velocity space,another algorithm is dynamic window method(DWA)which satisfies the USV kinetic constraints.The new method is optimized by the delineation of the absolute barrier area and the improvement of the evaluation function.Finally,the simulation experiment is carried out under different threat scenarios through MATLAB.The simulation results show that the improved threat avoidance method can guide the USV to avoid the danger situation in the dynamic and complicated environment.