仓储机器人系统多机协作规划问题研究-控制工程专业论文.docxVIP

万方数据 万方数据 西南科技大学硕士研究生学位论文第 西南科技大学硕士研究生学位论文 第 I 页 摘 要 仓储多机器人系统作为智能化仓储的重要部分，包含了多机任务分配、 路径规划和拥堵协调等技术，随着机器人数量增加，系统的抗拥堵能力受 到挑战。本文针对多机器人系统中的任务分配、路径规划、路径拥堵等问 题，模拟典型二维空间仓储机器人控制的主要流程和方法，系统性的仿真 实现多机器人的协作规划问题。 首先选择集中式的多机器人控制体系结构，针对任务分配提出一种基 于状态表的市场机制任务分配方法，综合考虑了任务分配中各种因素的影 响力度，保证机器人具备一定的任务协调能力 ；引入三维时空地图解决二 维地图中路径冲突问题，并改进 A*算法，使机器人路径搜索具备方向性， 提升了 A*算法的时间和空间利用效率；在系统出现紊乱的情况下，机器人 可能受到外力干扰，需要靠自身存储的环境信息与动力系 统到达目标点， 本文提出了目标可达 性分析法，实时判断机器人是否可到达目标点。 课题通过开发抽象的二维仓储多机协作平台，综合测试多机器人系统 的抗拥堵性能，并验证了系统临时增加任务的可行性， 应用上述方法使系 统并行运行的机器人数量 得到提升。 关键词：多机器人系统 智能仓储 任务分配 路径规划 A*算法 西南科技大学硕士研究生学位论文第 西南科技大学硕士研究生学位论文 第 II 页 Abstract As an important part of Intelligent warehouse logistics, multi-robot system contains multi-agent task allocation, path planning and path conflict of multi-robot. With the increasing number of robots, the systems ability to resist the congestion is being challenged. In this paper, taking the problems of task allocation, and path conflicts, there simulated the process and the main method of the multi-agent system in warehousing. It selected the centralized multi-robot control system structure at first. In view of the task allocation in this paper, considering the influence of various factors in the task assignment, a market mechanism for task allocation method based state table was proposed which make sure the robot have a certain task coordination. It studied the A* algorithm based the three dimensions of space and time, through adding directional constraints to heuristic rules to enhance the efficiency of algorithm. In the case of the disorder of the system, the robot may be interfered by external force. There put forwards a method of feasible analysis for global path planning. We developed a multi-agent simulation platform which was based on two dimensional space. Take a comprehensive test multi-robot system performance, using the above methods to make number of robots which run in parallel get promoted, and verify the feasibility of system when it temporarily increase tasks.