基于群体智能优化的AGV路径规划算法研究

发布时间：2018-01-12 13:29

本文关键词：基于群体智能优化的AGV路径规划算法研究　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：自动导引车隶属于移动机器人范畴,随着社会的进步和工业技术的发展,AGV已经成为当今工业生产、自动化仓储系统中的重要工具之一,通过对AGV及其关键技术的研究可以降低生产成本,提高生产效率,具有重要现实意义。自动导引车的路径规划是整个系统的关键技术之一,其特点有非线性、复杂性、约束性等,多年来作为特殊机器人的自动导引车路径规划算法的研究方兴未艾,基于数学模型的传统算法如禁忌法、栅格法、人工势场法等难以取得理想的效果,均有鲁棒性差、精度差、效率差等问题;而基于对社会性昆虫行为的模拟产生的一系列全局寻优的群体智能优化算法如蚁群算法、猫群算法等,则具有鲁棒性强、全局寻优、并行性等特点。因此本论文主要工作是基于群体智能优化算法研究自动导引车的路径规划。本文根据AGV的应用需求,将AGV的路径规划分为单任务目标路径规划和多任务目标路径规划。通过对比诸多文献中将机器人路径规划模型转为旅行商问题的思路,分析多任务目标AGV小车的路径规划特点,将其转化为旅行商问题,并且提出了三种算法来解决AGV系统的路径规划。(1)首先通过分析差分进化算法求解旅行商问题的思路,提出了离散猫群算法来求解AGV路径规划,通过引入位置-次序编码,将猫群算法扩展到离散域,使之可以用于求解该问题,并进一步分析了猫群算法中的重要参数对算法性能的影响。(2)其次针对蚁群算法在求解旅行商问题存在局部寻优与收敛性的矛盾,提出两种改进蚁群算法求解AGV路径规划:其一基于差分进化算法的全局寻优性,引入多种群差分蚁群算法,通过对分组后的蚁群进行不同的差分进化方式,有效提高最优解的搜索概率;其二根据猫群算法搜索模式的特点,引入基于猫群搜索的蚁群算法,实现蚁群个体在当前解集周围的局部搜索,有效改善蚁群算法寻优性。最后,分析了多AGV系统路径中出现的冲突问题,针对在交叉冲突中,现有的解决方案存在算法复杂,参数过多等问题,提出一种简单可控的解决方案,即采用提前交叉和蚁群算法局部二次规划。基于蚁群算法对每辆AGV小车进行预路径规划,通过检测交叉栅格,判断是否出现冲突,如果出现冲突,对于优先级较低的小车采用人工交叉法,使得小车经过交叉路径时存在时间差,并进行蚁群局部二次路径规划。通过Matlab仿真实验表明,该方案可以用于解决单任务目标多AGV系统中的交叉冲突现象。
[Abstract]:Automatic guided vehicle belongs to the category of mobile robot. With the progress of society and the development of industrial technology, AGV has become one of the most important tools in industrial production and automatic warehousing system. The research of AGV and its key technology can reduce the production cost and improve the production efficiency. The path planning of the automatic guided vehicle is one of the key technologies of the whole system. Its characteristics are nonlinear, complexity, constraint and so on. As a special robot, the path planning algorithm of automatic guided vehicle is in the ascendant for many years. The traditional algorithms based on mathematical model, such as Tabu method and grid method, are in the ascendant. The artificial potential field method is difficult to achieve the ideal effect, which has the problems of poor robustness, poor precision and low efficiency. Based on the simulation of social insect behavior, a series of global optimization algorithms, such as ant colony algorithm, cat swarm optimization algorithm, have strong robustness and global optimization. Therefore, the main work of this paper is to study the path planning of automatic guided vehicle based on swarm intelligence optimization algorithm. This paper is based on the application requirements of AGV. The path planning of AGV is divided into single-task target path planning and multi-task objective path planning. By comparing the robot path planning model into the traveling salesman problem in many literatures. This paper analyzes the path planning characteristics of multi-task target AGV vehicle and transforms it into a traveling salesman problem. And three algorithms are proposed to solve the path planning of AGV system. Firstly, through the analysis of differential evolution algorithm to solve the traveling salesman problem, a discrete cat swarm algorithm is proposed to solve the AGV path planning. By introducing position-order coding, the cat swarm algorithm is extended to discrete domain, which can be used to solve the problem. Furthermore, the influence of important parameters on the performance of the algorithm is analyzed. Secondly, the ant colony algorithm has the contradiction of local optimization and convergence in solving the traveling salesman problem. Two improved ant colony algorithms are proposed to solve the AGV path planning. One is based on the global optimization of differential evolutionary algorithm, the introduction of multi-colony differential ant colony algorithm, through the grouping ant colony differential evolution of different ways. Improve the search probability of the optimal solution effectively; Secondly, according to the characteristics of the cat colony algorithm search pattern, the ant colony algorithm based on the cat colony search is introduced to realize the local search of the ant colony individual around the current solution set, so as to effectively improve the optimization of the ant colony algorithm. This paper analyzes the conflict problem in the path of multi-#en0# system, and proposes a simple and controllable solution to the problem of complex algorithm and excessive parameters in the existing solutions. Based on the ant colony algorithm, the pre-path planning of each AGV vehicle is carried out, and the crossover grid is detected to determine whether or not there is a conflict, if there is a conflict. Artificial crossover method is used for the vehicles with lower priority, which makes the cars have time difference when they cross the path, and carries out the local secondary path planning of ant colony. The simulation results of Matlab show that. This scheme can be used to solve the cross-conflict phenomenon in single-task multi-AGV system.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TP23

【相似文献】