Research On Cooperative Path Planning Method Of Multi-UAV Long-term Operation

Intelligent UAV is widely used in the scene of acquiring natural environment information.When UAV is collecting information,the situation is often faced with a variety of information and wide distribution of task points.In the operation under the condition,the time spent will lead to different flight routes vary greatly.Therefore,the optimal path planning for UAV is very important.In addition,the endurance time of UAV is very limited,how to improve the endurance time of UAV is also an urgent problem.Based on the background of UAV multi task long-term operation path planning,this paper studies single UAV multi task long-term operation,UAVs multi task long-term operation and UAVs multi task multi-objective long-term operation path planning.In order to solve the problem of UAV long-term operation path planning,on the Genetic Algorithm(GA)and Multi-Objective Evolutionary Algorithm based on Decomposition(MOEA/D)improvements to calculate the optimal path,and realize the UAV long-term operation.The specific research contents of this paper are as follows:1)In order to solve the problems such as short endurance time of single UAV、slow convergence speed and falling into local optimum easily of traditional GA for solving single UAV multi task long-time operation path planning problem,the fixed charging point problem is studied,and UAV path planning method based on improved GA is proposed.Firstly,the environment model of the working area is established;based on the study of UAV charging constraints,the fixed charging point problem is designed and transformed into Traveling Salesman Problem(TSP)by using graph theory method.Secondly,in the TSP solution,the three parts of GA selection,crossover,and mutation are improved respectively,and the cubic B-spline curve is used to ensure the smooth and flyable path;Finally,verify the improved GA in the design environment of this article,the experimental results show that the convergence speed and generation value of the improved GA are reduced by 26.2%and 37.5%compared with the traditional GA,and the convergence speed and generation value of the improved GA are reduced by 19.1%and 28.6%compared with the ant colony algorithm.At the same time,the algorithm is tested in different obstacles,different space size and different task point distribution environment,which proves that the algorithm can adapt to this kind of environment and meet the requirements of single UAV multi task long-term operation path planning.2)Aiming at the problem of large number of working points in the environment and insufficient capability of single UAV,research on multi-UAV multi-task long-term operation path planning.Firstly,a Multi-Traveling Salesman Problem(MTSP)is established,and the MTSP is converted into m TSP;Secondly,in order to solve the problem such as low evolution efficiency,high computational complexity,easy to fall into local optima and low success rate of Genetic Algorithm and Variable Neighborhood Search Cooperative Algorithm(GAVC),multi-UAV path planning algorithm based on Improved Adaptive Genetic-Variable Neighborhood Collaborative Search Algorithm(IAG-VCA).In the IAG-VCA,the four parts of population initialization,adaptive crosover rate,adaptive mutation rate and neighborhood structure are improved respectively,which makes the IAG-VCA can better solve the UAVs multi task long-term operation path planning.Finally,verify the IAG-VCA in the design environment of this article,IAG-VCA is tested.The results show that IAG-VCA can reduce the generation value and convergence rate by at least 8.4%and 33.3%compared with Adaptive Genetic Algorithm(AGA),Improved Adaptive Genetic Algorithm(IAGA)and GAVC,and in the environment designed in this paper,the search success rate is as high as 96.3%,which meets the requirement of UAVs multi-task and long-term operation path planning.3)For the multi-objective optimization problem,UAVs multi task multi-objective long-term operation path planning is studied.Firstly,establish a multi-objective problem model,and optimize the path length and time of the UAVs at the same time;Secondly,aiming at the shortcomings of poor convergence and poor diversity in the optimal set of MOEA/D in the process of solving this problem,Based on Adaptive Loop Optimization Strategy MOEA/D(MOEA/D-ALOS)is proposed.In MOEA/D-ALOS,the evolutionary operator and neighborhood substitution strategy are improved respectively to maintain the convergence of the algorithm and ensure its diversity;Finally,in the ZDT and DTLZ series function test,and combined with MOEA/D,second-generation Non-dominated Sorting Genetic Algorithm-Ⅱ(NSGA-Ⅱ),MOEA/D based on Distance Update strategy(MOEA/D-DU)and Multi-Objective Evolutionary Algorithm based on Decomposition and Coevolution(MOEA/DCE).The final results show that MOEA/D-ALOS has the best performance,the average set coverage is higher than 71.6%,and the super volume value is the best in six test functions,which is 29%higher than the suboptimal MOEA/DCE.It can complete more complex environment multiple tasks and multiple target path planning for UAV.