A UAV obstacle avoidance path planning approach based on improved multi-trial vector-based differential evolution
Changsheng Zhu, Yafeng Zhao, Hongwei Bai, Yujie Nan, Minrui Zhao · Measurement Science and Technology · 2025
Abstract To address the issues of search-based algorithms being prone to local optima, the technical immaturity of multi-strategy optimization algorithms in path planning, and the significant storage overhead and computational burden of traditional 3D raster maps as environmental complexity grows, an improved multi-trial vector-based differential evolution (IMTDE) method for 2D raster height-filled 3D path planning is proposed, which aims to optimize the energy saving and safety of flight paths. By utilizing control point guidance to direct the evolution of the path, the method ensures progress toward the globally optimal solution, thereby accelerating convergence, introducing multiple vector producers and an adaptive step-size Gaussian wandering strategy with a winner-based allocation strategy for subpopulations, and adding a lifetime profile sharing mechanism to facilitate information transfer. To evaluate the effectiveness of the algorithm, the proposed method is tested on standard benchmark functions and compared with six classical intelligent optimization algorithms introduced in recent years, and the results show that the algorithm’s comprehensiveness is better than the comparison algorithms on both single-peak and multi-peak functions. Simulated in complex terrain and obstacle environments, the IMTDE algorithm outperforms the comparison algorithm in terms of average path length, optimal path length, and worst path length in scenarios 1 and 2, and its average total cost is 9.85% and 9.7% less than that of the comparison algorithm, respectively. Therefore, the IMTDE algorithm shows significant advantages in path length, running time, and robustness, so the path of the unmanned aerial vehicle planned by IMTDE has lower energy consumption.