Limited by the insufficiency of single UAV’s load and flight time capabilities, the multi-UAV (unmanned aerial vehicle) collaboration to improve mission efficiency and expand mission functions has become the focus of current UAV theory and application research. In this paper, the research on UAV global path planning is carried out using the ant colony algorithm, and an indoor UAV path planning model based on the ant colony algorithm is constructed. In order to improve the efficiency of the algorithm, enhance the adaptability and robustness of the algorithm, a distributed path planning algorithm based on the dual decomposition UAV communication chain is proposed. This algorithm improves the basic ant colony algorithm from the aspects of path selection, pheromone update, and rollback strategy in view of the inherent shortcomings of the ant colony algorithm. In order to achieve the best performance of the algorithm, this paper analyzes each parameter in the ant colony algorithm in depth and obtains the optimal combination of parameters. The construction method of the Voronoi diagram was improved, and the method was simulated to verify that the method can obtain a Voronoi diagram path that is safer than the original method under certain time conditions. Through the principle analysis and simulation verification of the Dijkstra algorithm and the dual decomposition ant colony algorithm, it is concluded that the dual decomposition ant colony algorithm is more efficient in pathfinding. Finally, through simulation, it was verified that the dual decomposition ant colony algorithm can plan a safe and reasonable flight path for multiple UAV formation flights in an offline state and achieve offline global obstacle avoidance for multiple UAVs.

中文翻译：

---

基于对偶分解的无人机通信链分布式路径规划

受限于单架无人机的载荷和飞行时间能力不足，多无人机协同提高任务效率、扩展任务功能成为当前无人机理论和应用研究的重点。本文利用蚁群算法进行无人机全局路径规划研究，构建了基于蚁群算法的室内无人机路径规划模型。为了提高算法的效率，增强算法的适应性和鲁棒性，提出了一种基于对偶分解无人机通信链的分布式路径规划算法。该算法从路径选择、信息素更新、针对蚁群算法固有的缺点，提出回滚策略。为了达到算法的最佳性能，本文对蚁群算法中的各个参数进行了深入分析，得到了最优的参数组合。改进了Voronoi图的构造方法，并对该方法进行了仿真，验证了该方法在一定时间条件下可以获得比原方法更安全的Voronoi图路径。通过对Dijkstra算法和对偶分解蚁群算法的原理分析和仿真验证，得出对偶分解蚁群算法的寻路效率更高。最后通过仿真，