复杂环境下移动机器人路径规划方法研究

发布时间：2018-05-16 08:03

本文选题：移动机器人 + 路径规划　；参考：《天津理工大学》2017年硕士论文

【摘要】：路径规划问题是移动机器人研究领域的一个最基本、最关键的课题,它解决的是移动机器人在工作环境中如何行走的问题。本文主要研究移动机器人在不同环境下的路径规划,包括一般环境下点对点之间的路径规划与改善,温室环境下围绕农作物生长区域半包围式的路径规划和三维立体环境下的路径规划。针对在这三种环境下机器人路径规划的不同特点分别开发了相应的算法,并通过仿真实验验证了相应算法的有效性。本文的研究工作主要分为以下三个部分:1.提出用于改善路径质量的平滑蚁群算法。在利用蚁群算法进行路径规划时,机器人所走路径节点都是自由栅格的中心点,这样就造成了机器人会出现不必要的转弯和行走一些多余的路径。应用平滑蚁群算法将蚁群算法规划出来的最优路径进行平滑处理,在处理过程中,将机器人所在的当前节点与转折点后的其它节点连接,如果连接线没有穿越障碍物区域,则当前连接的线段就可以作为新的路径代替原来路径,并将中间不必要的节点删除,否则对当前路径不作任何改动。2.提出一种温室环境下的路径规划方法。在温室环境下,农民会将农作物按照一定结构来种植,机器人在温室内工作,就需要围绕着农作物进行运动,运动的同时对农作物进行喷药、施肥等工作。针对温室内机器人的特殊路径规划方式,将蚁群算法和人工势场法相结合,综合利用蚁群算法全局搜索和正反馈的特点以及人工势场法能够定向运动和预先规避障碍物的优点,在规定机器人的工作区域和起止点之后,新的势场蚁群算法会根据要求规划出机器人的最优路径。3.提出一种在三维环境下利用蚁群算法进行路径规划的方法。在构建三维地图时,首先将移动机器人所处的三维环境划分为一个个的平面,然后再将这些平面进行栅格化。在利用蚁群算法进行路径规划的过程中,将蚂蚁释放的信息素全部存储在路径节点上,而不是放在蚂蚁的行走路径中,这样能大大减少信息素的存储空间,同时构建了新的期望启发函数,提高了路径搜索效率。
[Abstract]:Path planning is one of the most basic and critical topics in the field of mobile robot research. It solves the problem of how mobile robot walks in the working environment. This paper mainly studies the path planning of mobile robot in different environments, including the path planning and improvement between point to point in general environment. Semi-bounding path planning around crop growing area in greenhouse environment and path planning in three-dimensional environment. According to the different characteristics of robot path planning in these three environments, the corresponding algorithms are developed, and the effectiveness of the corresponding algorithms is verified by simulation experiments. The research work of this paper is divided into the following three parts: 1. A smoothing ant colony algorithm for improving path quality is proposed. When the ant colony algorithm is used in path planning, the path nodes taken by the robot are all the center points of the free grid, which causes the robot to make unnecessary turns and walk some redundant paths. The smooth ant colony algorithm is used to smooth the optimal path of the ant colony algorithm. In the process of processing, the current node of the robot is connected with the other nodes after the turning point, if the connecting line does not cross the obstacle area, Then the line segment of the current connection can replace the original path as a new path, and remove unnecessary nodes from the middle, otherwise, no changes to the current path. 2. A path planning method in greenhouse environment is proposed. In the greenhouse environment, farmers will plant crops according to a certain structure, robot work in the greenhouse, it is necessary to move around the crops, while the movement of crops spraying, fertilization and other work. Aiming at the special path planning method of robot in greenhouse, this paper combines ant colony algorithm and artificial potential field method, synthesizes the characteristics of global search and positive feedback of ant colony algorithm and the advantages of artificial potential field method which can orient movement and avoid obstacles in advance. After defining the working area and starting and ending points of the robot, the new ant colony algorithm of potential field will plan the optimal path of the robot according to the requirement. An ant colony algorithm for path planning in three-dimensional environment is proposed. In the construction of 3D map, the 3D environment of the mobile robot is first divided into one plane, and then these planes are rasterized. In the course of path planning using ant colony algorithm, the pheromone released by ant is stored on the path node instead of in the walking path of ant, which can greatly reduce the storage space of pheromone. At the same time, a new expected heuristic function is constructed to improve the efficiency of path search.
【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】