Research On Task Decision-making For Multi-UAV Cooperative Reconnaissance

Under the complex battlefield environment in the future,how to use limited load resources to detect the maximum intelligence information will be an important basis for commanders' decision-making.In recent years,with the rapid development of UAV and the continuous improvement of new reconnaissance load technology,UAV is regarded as the main means of reconnaissance,and its task decision-making problem has already become one of the hot issues in the research of domestic and foreign scholars.So this dissertation mainly studies the problem of task decision-making for multi-UAV cooperative reconnaissance.Since multi-UAV cooperative reconnaissance task decision-making problem is not only a complicated multi-objective optimization and decision problem with multi-constraint,but also a large-scale coordination control problem,the dissertation first studies single UAV task decision-making problem,and then the research of multi-UAV task decision-making problem is carried out.This research can effectively work out the task set and reconnaissance order of each UAV with its limited load resource,making the total reconnaissance reward maximum.The main work and innovation points of this dissertation are as follows:(1)UAV reconnaissance revenue maximization approach based on improved cuckoo search(ICS)algorithm is presented.The method is mainly to get the reconnaissance time on each task of the UAV who has the ability to reconnoiter this task,which can make the total reconnaissance revenue maximum.First of all,information certainty index is raised to measure the reconnaissance revenue and then the mathematical model is established.Next,according to the basic principle of the emerging heuristic cuckoo search algorithm(CS),adaptive scaling factor is proposed to improve CS.Simulation results show that the improved CS algorithm can obtain the feasible solution in a shorter time,compared with basic CS algorithm.(2)UAV route planning approach based on improved discrete cuckoo search(IDCS)algorithm is proposed.Single UAV multi-reconnaissance tasks route planning problem is studied using the shortest total flight distance as objective function.At first,the problem is simplified as the standard travelling salesman problem(TSP)and the mathematical model considering various constraints are established.Then an improved discrete cuckoo search algorithm is put forward solving the route planning problem.Simulations prove that the improved DCS can effectively solve the UAV route planning problem in a shorter time,at the same time,at the same time,the algorithm convergence is better.(3)Multi-UAV cooperative task allocation approach based on the consensus-based bundle algorithm(CBBA)is presented.At the first,based on the characteristics of UAV and task area resources,“resources-demand” matrix is put forward.Next,according to the purpose of the task allocation,this dissertation puts forward the UAV arrival time index and the "UAV-task" time index.Then the multi-UAV cooperative task allocation model is set up.At last,in view of the complexity of the problem and the weaknesses of the centralized algorithms,the extension consensus-based bundle algorithm(CBBA)is chosen to solve this problem.The simulation results show that ECBBA can obtain the maximum objective function value and get each UAV's task set and corresponding reconnaissance order in a much shorter time.