关键词: 蚁群-聚类算法, 动态路径规划, 自适应半径, 对角障碍, 平滑机制

Abstract: When the traditional ant colony algorithm is applied to the path planning of robot in the time-varying environments, there are several problems such as slower convergence speed, lager path integral turning angle and poor self-adaptive ability. So, a self-adjustable dynamic path plan algorithm is presented based on both the ant colony and the clustering algorithm in this paper. In the new ant colony-clustering algorithm proposed here, a suitable searching radius is chosen automatically by using the clustering algorithm for different environmental complexities, which can make full use of the limited computing resources of the robot and improve the convergence speed. By identifying the diagonal obstacles with the proposed virtual obstacle generation method, virtual obstacles are used to guarantee that the problem of planning path crossing diagonal obstacles is nonexistent. The dynamic searching path is smoothed and optimized by the smoothing mechanism, which can effectively reduce the path length and the cumulative turning angle. Simulation examples are provided to show that the algorithm can find suitable searching radius according to different obstacle distributions, complete the dynamic path planning, show good environmental self-adaptability and better total path optimization performances.

Key words: ant colony-clustering algorithm, dynamic path planning, self-adjusting radius, diagonal obstacle, smoothing mechanism