基于改进遗传算法的移动机器人路径规划研究

发布时间：2018-07-23 07:43

【摘要】：移动机器人集中了传感器技术、机械、电子、计算机、自动化控制以及人工智能等多学科的研究成果,其发展是一个国家高科技水平和工业自动化程度的重要体现。在移动机器人相关的技术研究中,路径规划技术是机器人研究领域的重要组成部分,也是机器人完成指定任务的必要基础和根本保障,其任务是在一个存在障碍物的有界工作空间内,依据相应的评价标准(移动时间、路径长度、能量消耗等)能够根据周围环境信息自主的从给定起点运动到目标点,同时确保机器人与障碍物之间、机器人与机器人之间不发生碰撞。至今,大部分研究都集中在静态环境下单机器人路径规划,而多机器人系统能够有效的处理复杂、动态性高以及需要并行完成的任务,具有单机器人无法比拟的优势,因此研究动态以及多机器人环境下的路径规划问题具有重要意义。目前,现有的优化算法在求解路径规划问题时都存在各自的缺陷,因此寻求更佳的算法就成为该领域的一个研究重点。鉴于遗传算法具有较强的鲁棒性、并行性以及较强的全局搜索能力等优点,本文研究设计了改进的遗传算法,并将其应用于单机器人以及多机器人路径规划中。论文的主要研究内容如下:1.研究了全局环境下移动机器人路径规划问题,为了解决基本遗传算法在求解机器人路径规划问题中存在的收敛速度慢、易陷入局部最优等不足,对遗传算法进行了改进。引入了人工势场法进行种群初始化,提出了一种基于种群多样性程度评价的自适应选择方法,设计了自适应交叉和变异概率,改善了基本遗传算法收敛速度慢和早熟现象,提高了算法的求解质量。在栅格环境下进行多次仿真实验,证明了所提出改进遗传算法的可行性和有效性。2.针对动态环境下单移动机器人路径规划问题,在规划过程中将全局路径规划与局部路径规划相结合,并且根据机器人与动态障碍物碰撞类型的不同,制定了有效的避碰策略。仿真实验表明该方法能够有效的指导机器人在动态环境中实现避障,获得无碰最优或次优路径。3.针对动态环境下多机器人路径规划进行了研究,为解决机器人间的路径冲突问题,制定了有效的路径协调策略,实验结果表明该方法能够较好地实现多机器人路径规划。
[Abstract]:The development of mobile robot, which includes sensor technology, machinery, electronics, computer, automation control and artificial intelligence, is an important embodiment of a country's high-tech level and industrial automation. In the research of mobile robot technology, path planning technology is an important part of robot research field, and it is also the necessary foundation and fundamental guarantee for robot to complete the assigned task. In a bounded workspace with obstacles, according to the corresponding evaluation criteria (moving time, path length, energy consumption, etc.), the task is to automatically move from the given starting point to the target point according to the information of the surrounding environment. At the same time, make sure that there is no collision between robot and obstacle and between robot and robot. Up to now, most of the researches have focused on the path planning of the robot in static environment, and the multi-robot system can deal with complex, dynamic and parallel tasks effectively, which has the advantage that the single robot can't compare with each other. Therefore, it is of great significance to study the dynamic and path planning problems in multi-robot environment. At present, the existing optimization algorithms have their own defects in solving the path planning problem, so the search for better algorithms has become a research focus in this field. In view of the strong robustness, parallelism and global search ability of genetic algorithm, an improved genetic algorithm is designed and applied to single robot and multi-robot path planning. The main contents of this thesis are as follows: 1. In this paper, the path planning problem of mobile robot in the global environment is studied. In order to solve the shortcomings of the basic genetic algorithm in solving the robot path planning problems, such as slow convergence speed and easy to fall into local optimum, the genetic algorithm is improved. The artificial potential field method is introduced to initialize the population, and an adaptive selection method based on the evaluation of population diversity is proposed. The adaptive crossover and mutation probability are designed to improve the slow convergence speed and premature convergence of the basic genetic algorithm. The quality of the algorithm is improved. Several simulation experiments in grid environment show that the proposed improved genetic algorithm is feasible and effective. For the path planning problem of mobile robot in dynamic environment, the global path planning and local path planning are combined in the planning process, and an effective collision avoidance strategy is proposed according to the different collision types between the robot and the dynamic obstacle. The simulation results show that this method can effectively guide the robot to avoid obstacles in dynamic environment and obtain the optimal or sub-optimal path. 3. In this paper, the path planning of multi-robot in dynamic environment is studied. In order to solve the problem of path conflict between robots, an effective path coordination strategy is proposed. The experimental results show that the method can achieve multi-robot path planning well.
【学位授予单位】：安徽工程大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242;TP18

【参考文献】