Reasearch And Implementation Of Multi-UAVs Cooperative Task Planning

Along with the continuous progress of science and technology in various countries,the situation of modern battlefield is becoming more and more complex,then the advantages of UAV in modern war are fully reflected.With the increasing difficulty of UAV in the battlefield,the Multi-UAVs operation mode gradually replaces the previous single UAV operation mode,so the ability of Multi-UAVs cooperative task planning needs to be further upgraded.This topic mainly studies the problem of Multi-UAVs cooperative task planning,mainly for two parts,the first part is Multi-UAVs task assignment,reasonable task allocation is the guarantee of multi UAV cooperative task planning,the second part is Multi-UAVs track planning,high-quality track planning is the core content of Multi-UAVs cooperative task planning.The branch and bound algorithm is proposed to solve the Multi-UAVs task assignment problem.According to the needs of modern battlefield,in order to better complete the task,Multi-UAVs groups will have many different types of UAVs.Due to the differences in the attributes of each UAV,it becomes an assignment problem to select the right UAV to complete tasks according to the different needs of tasks.The branch and bound algorithm can find the feasible solution of task assignment by expanding the branch continuously and control the search direction of branch by delimiting the boundary of feasible solution,then find the optimal solution of task assignment problem.Aiming at the problem of Multi-UAVs track planning,this paper puts forward the obstacle avoidance strategy of track planning according to the problems that maybe encountered in the process of Multi-UAVs track planning,which respectively detects circular,polygonal and rectangular obstacles.For circular and polygonal obstacles,this paper treats them as the detection of circular obstacles by linearization.For the volume problem of UAV,this paper deals with the expansion of obstacles,the obstacle avoidance strategy is that the path nodes generated by track planning should not be in the external square of circular obstacles.For rectangular obstacles,this paper judges whether the line between the path nodes generated in the track planning process and the adjacent nodes intersects with any side of the rectangular obstacles,so as to complete the collision detection task of rectangular obstacles and transform the obstacle avoidance strategy into the intersection problem of line and rectangle.This paper solves the obstacle avoidance problem in the course of UAV track planning by these two methods.Aiming at the Multi-UAVs track planning problem,this paper proposes a fast search random tree algorithm(RRT algorithm).RRT algorithm is a kind of random sampling algorithm with incremental growth,in which the growth of random number is the key step,including the strategy selection of task tree searching from a certain point to the surrounding and the determination of the exploration distance to the target point.In order to solve this problem,this paper tests and compares the probability parameters and step parameters of RRT algorithm,then selects the appropriate parameters to solve the problem of track planning problem.Aiming at the RRT algorithm that has been put forward,in order to get a better track route,this paper optimizes and improves this algorithm.The improved algorithm first obtains a track route that meets the task requirements through RRT algorithm,then reselects the parent node for the existing node and redistributes the random tree line to obtain a track route that is better than the original one.After the simulation experiment of RRT algorithm and the improved RRT algorithm,through the comparison of the experimental results,this paper concludesthat when generating the track route,the improved RRT algorithm generates fewer turning points of the track route and the track route is more smooth,that is to say,the improved RRT algorithm generates better track route than the former algorithm.