Multi-Strategy Evolutionary Mechanism for UAV 3D Path Planning in Multi-Obstacle Environments

Aiming at the problems such as low convergence accuracy and insufficient stability in path planning for unmanned aerial vehicles (UAVs) in 3D multi-obstacle scenarios, a multi-strategy evolutionary particle swarm optimization (MSEPSO) algorithm is proposed. In the initialization stage, aiming at the problem that particle swarm optimization is sensitive to the initial position of particles, the initial distribution of particles is optimized through Latin hypercube sampling to improve the population diversity; During the evolutionary stage, a “balance-memory-enhancement” evolutionary framework is designed, which utilizes a nonlinear iterative strategy to balance global development and local search. The personal history memory mechanism is adopted to enhance the global exploitation ability of the algorithm. Evolutionary particles are introduced to enhance the exploration ability of the population in the vicinity of the group＇s extreme values, reducing the probability of the algorithm getting stuck in local optima. Experimental results from comparisons on the CEC2020 test function set and in mountain/urban scenarios demonstrate that MSEPSO exhibits stable optimization performance, enabling the planning of safer paths with shorter lengths and higher smoothness.