Research On UAV Path Planning For OSG Visual Simulation Platform

At present,the application of UAV is becoming more and more widespread.Complex task and environment put forward high requirements on the performance of UAV.The research of UAV path planning is of great significance.The UAV path planning researches mainly focus on single algorithms,but there are certain limitations of single algorithms.While the multiple fusion algorithm with better performance is less studied,as well as the inconformity between path planning trajectory and actual UAV flight of previous researches.An Ant Colony Optimization based on mixed heuristic information was proposed to overcome local optimization and poor convergence of traditional ant colony algorithm.In this algorithm,mixed heuristic information containing cost function and bending suppressor of A* algorithm was set up in traditional Ant Colony Optimization,the pheromone update mechanism was adjusted,the maximum and minimum pheromone values on the path were limited,so as to increase the convergence speed and reduce the number of bending.Path fusion algorithm based on Rapidly-exploring Random Tree(RRT)global search was proposed to overcome the inconformity between RRT solution path and actual flight requirements of UAV.Global path planning was carried out based on RRT algorithm,then the RRT algorithm solution path was smoothed and optimized by Bezier curve to obtain the ultimate planned path.A 3D visualization system was designed and built based on Open Scene Graph(OSG).The system has a friendly interactive interface which is easy to operate.In the system,the UAV model is simulated for path planning,and the effectiveness of the proposed path planning algorithm is verified.The simulation results verified that the convergence efficiency and bending suppression effect of ant colony algorithm based on mixed heuristic information were greatly improved.In a complex simulation map environments,the convergence efficiency was improved by about 40% and the path bending number was reduced by about 25%.The planned path trajectory of the path fusion algorithm based on RRT global search was optimized,and the planned path was shortened by about 2.5% in various map environments without bending angles,which met the actual flight requirements of the UAV.