3D Flight Path Planning for UAV Based on Improved Particle Swarm Optimization Algorithm

Unmanned aerial vehicle (UAV) has been widely used in various fields such as agriculture, petroleum, military, meteorology, and geographic surveying. In actual flight, unmanned aerial vehicle needs to find the shortest path and avoid all threats. An improved particle swarm optimization algorithm combined with genetic algorithm method is presented in this paper to solve the path planning problem of UAV which easily falls into local optimization. The algorithm enhances early stage global search capability and later period local optimization capability compared to traditional particle swarm algorithm. To ensure that particles are distributed in key search areas of the environment, Gaussian distribution is used to initialize the particle distribution. Optimization ability can be further improved by the linear transformation operation on the well performing particles. Logistic function is employed to set the dynamic mutation probability. To improve the global search capability, random initialization operations are performed on particles with poor performance. Simulations in simple terrain and complex terrain environment are carried out to testify the feasibility and efficiency of the proposed improved algorithm.