Smooth Path Planning of Ackerman Chassis Robot based on Improved ant Colony Algorithm
Smooth Path Planning of Ackerman Chassis Robot based on Improved ant Colony Algorithm
复制标题
基于改进蚁群算法的阿克曼底盘机器人平滑路径规划
DOI：
10.46300/9106.2020.14.49
复制
发表时间：
2020-07
影响因子：
--
通讯作者：
Yili Zheng
中科院分区：
文献类型：
--
作者：
Guannan Lei;Yili Zheng
在机器人和自动驾驶领域，车辆无碰撞运动的自动路径规划是导航层面的一项重要任务。发现传统的路径规划算法和折线路径不能完全满足阿克曼底盘机器人的驱动要求。为了解决结构化环境下阿克曼底盘移动机器人的自主导航问题，提出一种新的改进算法。配置空间方法可以将机器人自身的结构尺寸参数引入到算法中。通过凸多边形检测方法，将地图中的局部U形区域转化为封闭区域。这两种策略的本质都是对地图进行预处理。初始信息素分布不再是全局均匀的，而是根据地形分布。结合泊松分布规律，将信息素的挥发因子由静态常数变为动态常数。该策略使得改进的信息素分布规律不仅避免了算法初期的随机性和盲目性，而且保证了算法中期蚁群的探索行为和引导作用。路径平滑也是一项具有挑战性的任务。该算法通过改进评价函数、去除冗余节点和2-turn算法来逐步优化路径。从而获得适合阿克曼机器人的无碰撞平滑路径。本文结合多种算法改进策略，不仅提高了蚁群算法路径探索的性能，而且规划了一条适合阿克曼移动机器人跟踪的平滑曲线路径而不是折线。通过MATLAB对算法进行编码和仿真，验证了算法的可行性和有效性。这将为后续算法移植提供重要依据，为阿克曼底盘机器人的路径跟踪控制奠定基础。
In the domain of robotics and autonomous driving, the automatic path planning of vehicle collision-free motion is an essential task on the navigation level. It is found that the traditional path planning algorithm and the ployline path cannot fully meet the driving requirements of Ackerman chassis robot. In order to solve the autonomous navigation problem of Ackerman chassis mobile robot in structured environment, this paper presents a new improved algorithm. The method of configuration space can introduce the robot's own structural size parameters into the algorithm. Through convex polygon detection method, the local U-shaped area in the map is transformed into a closed area. The essence of these two strategies is to preprocess the map. The initial pheromone distribution is no longer globally uniform, but is distributed according to the terrain. The volatilization factor of pheromone is changed from static constant to dynamic one, which is combined with Poisson distribution law. This strategy makes the improved pheromone distribution law not only avoid the randomness and blindness in the initial stage of the algorithm, but also ensure the ant colony's exploration behavior and guiding role in the middle stage of the algorithm. Path smoothing is also a challenging task. This algorithm optimizes the path step by step by improving the evaluation function, removing redundant nodes and 2-turning algorithm. Thus, a collision free smooth path suitable for Ackerman robot is obtained. This paper combines a variety of algorithm improvement strategies, not only improving the performance of ant colony algorithm path exploration, but also planning a smooth curve path suitable rather than polyline for Ackerman mobile robot tracking. The algorithm is coded and simulated by MATLAB, and the feasibility and effectiveness of the algorithm are verified. This will provide an important basis for the subsequent algorithm migration and lay the foundation for the path tracking control of the Ackermann chassis robot.