Unmanned Aerial Vehicle Trajectory Planning Technology Based on Improved ACO Fusion Algorithm
Hao Chen, Lankang Zhu, Fa Wang, Yue Yao, Wei Han · Unmanned Systems · 2025
With the advancement of contemporary science and technology, artificial intelligence has become a common practice across numerous industries, with unmanned aerial vehicles (UAVs) serving as its primary component. However, the current unmanned aerial vehicle trajectory planning is mostly limited to a single algorithm or a two-dimensional environment, with limited planning capability in a three-dimensional environment. Therefore, the study is to improve the ant colony algorithm and the artificial potential field method, respectively, and to combine the two improved algorithms to propose a fusion model, according to which the improved UAV trajectory planning (UAV-TP) system is designed. The outcomes indicated that the fusion model had the optimal mean and standard deviation on three different types of benchmark test functions, and possessed the best adaptability and stability. Applying the fusion algorithm to the experiments of the UAV-TP system, its optimal path in a simple 2D grid environment passed through 38 grids, produced 16 inflection points, 0 unsafe points, and searched for the optimal path at the 13th iteration. The average trajectory length in 20 experiments in a complex 3D environment was 27.59[Formula: see text]km, which was 28.43% and 16.21% lower than the rest of the comparison algorithms, respectively, and the average number of corners was seven. The outcomes reveal that the fusion algorithm proposed by the institute performs well in both trajectory planning and safe navigation, and can realize efficient navigation in three-dimensional complex environments.