【作者】 郭国强；

【导师】 巩云鹏；

【作者基本信息】 东北大学 ， 机械设计及理论， 2008， 硕士

【摘要】 在经济全球化和信息化的推动下,现代物流技术已经成为一个包含机械学、计算机科学、管理工程学和自动控制技术等多学科融合的技术,自动导引小车(AGV)正是现代物流技术的典型代表。作为物流配送过程中的关键问题之一,车辆路径规划问题(VRP)是指AGV按照某一性能指标(如距离、时间、能量等)搜索一条从起始点到目标点的最优或次优安全路径,因此它主要涉及的问题包括：如何利用已知AGV工作环境信息建立一个合理的模型,利用某种算法寻找一条从起始点到目标点的最优或者次优安全路径。本文首先对AGV的全局路径规划方法进行了研究。在工作环境已知的情况下,按照一定简化规则得到AGV运行环境的电子地图数据,从而对运行环境进行建模,再利用voronoi图对环境进行划分,最后采用dijkstra最优搜索算法为AGV寻找到一条威胁代价最小情况下的粗略最短路径。考虑到AGV的动力学约束,利用B样条曲线对粗略最短路径进行平滑和修正,得到最终全局优化路径。提出在利用二维仿真对路径规划算法进行理论验证的基础上,利用视景仿真(Scene Simulation)技术,模拟AGV运行的整个场景和AGV随时间变化的运行状态,从而对先期的理论算法进行检验和评估。在三维视景仿真软件Vega中,利用Hermite曲线对全局优化路径上的关键点进行拟合,构建平滑的曲线,从而根据曲线上点的位置坐标和欧拉角(x,y,z,h,p,r)来控制AGV的连续运动。对于路径规划与视景仿真的交互方法,本文对数据接口的定义,特殊效果、声效的模拟,人机交互的方法分别进行了详细的阐述,并对将Opengl引入到Vega开发环境的一些关键技术和方法进行了相关的研究。最后介绍了自动导引小车(AGV)运动视景仿真系统的体系结构和视景仿真的实验效果。仿真结果表明：自动导引小车(AGV)在按照全局路径规划求解得到的路径运动过程中,没有与障碍物等空间约束发生碰撞等事故,从而验证了全局路径规划算法的正确性。因此,利用视景仿真系统,能够使设计人员在虚拟场景中得到形象、直观、准确的信息,对基于voronoi图的路径规划理论和算法进行检验,从而提升了理论和算法在现实中的工程应用价值。更多还原

【Abstract】 With the impromotion of the economic globalization and information, the modern logistics has combined many other technologies, such as mechanics、computer science、manage engineering、automation technology etc. Automation Guided Vehicle (AGV) is just the representation of modern logistics technology. As one of the most important problems of the logistics process, Vehicle Routing Problem (VRP) is searching an optimal path without touch from the start to the end according to the performance index (such as:distance、time、energy) for AGV. So it includes these problems:how to build a reasonable model using the AGV’s working environment information that known, how to search an optimal path without touch from the start to the end if we could find the arithmetic.This thesis first focuses on the study of global path planning for AGV. Under the condition that we have known the AGV’s environment information, we can get the electronic map data according to some simplified regulations, and then we can use the data to model the environment. Using the voronoi diagram to divide the environment and the optimal searching arithmetic of dijkstra can be used to find an initial shortest path for AGV, at the same time, the threat weight of this way is the least. Considering the AGV’s dynamic constrains, the initial shortest path is smoothed by using B spline, then the latest global optimal path is gained.Using the two-dimensional simulation to verify the path planning arithmetic.On the basis of this, we use the scene simulation technology to simulate the AGV’s moving environment and it’s moving state according to time. All the methods are designed to illustrate and demonstrate the feasibility of the presented theory and arithmetic.In the 3D scene simulation software of Vega, the points from the global optimal path are fitting into Hermite spline. By Constructed the smooth spline, we can use the points’ positions and Ouler angles to control the AGV’s continual movement.As the method merges the path planning and the scene simulation, this thesis has a detail study on the next points:defining data interface, building special effects、Audio application, man-machine alternation method. And this thesis expatiate some key technology and development means about integration of OpenGL and Vega. At last, this thesis introduces the system’s structure of AGV movement scene simulation. The simulation result shows, in the course of AGV’s movement according to the path which calculated by the global path planning arithmetic, there is no collsion accident etc. between AGV and space restriction like obstacles. It proves the correctness of the global path planning arithmetic. With the scene simulation system, it can offers designer visual、intuitionistic and xact information in the visual environment, verifies the path planning theory and arithmetic based on voronoi graph. It advances the theory and arithmetic’s usefulness in reality.更多还原