Research On 3D UAV Path Planning Based On The Improved Algorithm

As an efficient aerial platform,unmanned aerial vehicle(UAV)has been widely used in various fields due to its high mobility and flexible operation.With the wide application of UAV,the airspace environment becomes increasingly tense.How to improve the autonomous flight capability of UAV has gradually become a key research field.As one of the key technologies in this field,path planning algorithm helps UAV to complete the task successfully.In this paper,based on the traditional classical algorithm and modern intelligent algorithm,modified UAV path planning algorithms are proposed to solve the problems such as the slow convergence speed and the path can hardly meet the constraints of UAV.The main research contents of this paper are as follows:Firstly,the basic principle and the improvements of rapidly-exploring random tree(RRT)algorithm are introduced.The advantages and disadvantages of different modified RRT algorithms are analyzed.Aiming at the shortcomings of RRT algorithm,the dynamic step BI-RRT algorithm is proposed,which improves the efficiency of RRT algorithm node exploration by introducing half angle tree clipping strategy and target-oriented sampling strategy.A dynamic step optimization strategy is proposed to solve the slow growth of random trees.The greedy algorithm is used to shorten the path distance.B-spline function is used to smooth the path.This paper continues to study rapidly-exploring random tree star(RRT*)algorithm with asymptotic optimality.The research shows that the convergence speed of RRT* algorithm is slow.In order to solve this problem,this paper proposes BPIB-RRT* algorithm based on bi-directional bias sampling heuristic function(Bi-bias())and bi-directional artificial potential field heuristic function(BPF())to improve the convergence speed of RRT* algorithm.Among them,Bi-bias()heuristic function biases the sampling points to make them closer to the direction of the target tree;BPF()heuristic function based on artificial potential field method sets the latest node on the target tree as the attraction pole to guide the rapid growth of the random tree.Secondly,this paper further studies and analyzes A* algorithm with cost estimation function.However,the convergence speed of A* algorithm is slow.To solve this problem,this paper proposes the bi-directional adaptive A* algorithm.Among them,the directional search strategy and the modified cost estimation function improve the expansion efficiency of A* algorithm and the path quality;the adaptive step strategy and adaptive weight strategy ensure that the algorithm converges to the optimal solution quickly;the re-wiring process reduces the distance cost.Finally,aiming at the path planning of multiple UAVs,this paper proposes global search and local search moth-flame algorithm(GLMFO)based on moth-flame algorithm.According to the constraints of multiple UAVs,the fitness function of the algorithm is modified to ensure the safety and optimality of the path;the population initialization strategy based on chaos and population diversity optimization strategy are proposed to ensure the diversity of the population;the adaptive weighted moth position update strategy is proposed to accelerate the convergence of the algorithm.