Abstract: An improved ant-colony algorithm was proposed to correct the navigation error in the path planning of UAVs under complex environments. Based on ant colony algorithm, firstly, the fitness of particle swarm optimization algorithm was introduced into pheromone update as heuristic value in the proposed algorithm, which improved the original pheromone updating model. Besides, a greedy strategy was used to carry out pruning operations when selecting the correction points, which optimized the selection strategy of the algorithm. Finally, the A* algorithm was used to replace the random initialization process of the original algorithm, which modified the update method of pheromone and optimized the quality of the generated paths. For the planned path, the Dubins curve was used to smooth the track. The smoothed track could not only meet the track constraints, but also meet the performance constraints of the aircraft. The research results show that: in parameter setting, when the number of ant colony is large, setting smaller heuristic constant can get better results. A* algorithm is used to initialize the greedy ant colony algorithm, which can effectively accelerate the convergence speed of ant colony algorithm and improve the quality of solution. Experiments show that the improved algorithm reduces the track length by about 6% and the time by about 25%.

Key words: traffic and transportation engineering; ant-colony algorithm; correction point; track planning; greedy strategy; UAV

摘要： 针对复杂环境下无人机航迹规划中航行误差的校正问题，提出一种改进的蚁群算法。该算法在蚁群算法的基础上，首先将粒子群算法中的适应度作为启发值引入信息素更新中，改进了原始的信息素更新模型；其次使用贪心策略在选择校正点时进行剪枝运算，优化了算法的选择策略；最后使用A*算法替代原始算法的随机初始化，修改了信息素的更新方式，优化了生成路径的质量。对规划路径，使用Dubins曲线对航迹进行光滑，光滑后航迹既能满足航迹约束，也能满足飞行器的性能约束。研究结果表明： 在参数设置上，当蚁群数量较大时，设置较小的启发值常数能获得更好的结果；对贪心蚁群算法使用A*算法进行初始化，能有效加速蚁群算法收敛速度，提高解的质量，实验显示改进后航迹长度减少了约6%，时间减少了约25%。

关键词: 交通运输工程；蚁群算法；校正点；航迹规划；贪心策略；无人机