移动视觉测量机器人任务规划方法研究

发布时间：2019-01-08 14:18

【摘要】：随着国防工业的不断发展以及科学技术水平的不断提高,大型航天产品的智能装配技术得以发展,移动视觉测量机器人在智能装配的过程中起到了关键性的作用。视觉测量和路径规划是移动视觉测量机器人任务规划系统中的两个重要的研究方向。对于待装配的被测物,我们需要确定机器人的一个最优观测位置,使其在该位置进行测量产生的误差最小,让机器人能够更好地指导装配过程。本文以单目视觉为测量方案,提出了一种机器人最优观测位置确定的方法。路径规划的目的一般是确定机器人在工作环境中移动的最短路径,本文对A*算法进行改进,提出了基于改进A*算法的路径规划方法。基于栅格法考虑环境中标志点、被测物和观测位置等因素提出了三种不同的地图构建方法。本文的主要研究工作如下:第一,对移动视觉测量机器人观测位置规划问题进行研究。建立单目视觉测量模型。根据单目视觉成像原理确定测量方案,以此为基础根据几何关系和坐标转换分别针对二维和三维情况建立视觉测量模型。针对不同的工况建立最优观测位置测量模型,确定最优观测位置。第二,对移动视觉测量机器人的路径规划方法进行研究。阐述A*算法的实现过程,基于A*算法在栅格地图中进行路径规划方法的仿真。针对该方法会出现的碰撞问题,改进路径规划方法,使其能够避免碰撞问题。第三,对移动视觉测量机器人的地图构建方法进行研究。与传统地图构建不同,本文的地图构建引入标志点和被测物等因素,并以此为基础,分别针对单独考虑标志点因素,综合考虑标志点和被测物因素,综合考虑标志点、被测物和观测位置因素三种不同的地图构建方法,利用基于改进A*算法的路径规划方法进行仿真,并分析对比不同方法的优缺点和应用场合。第四,通过实验验证最优观察位置测量模型和路径规划方法,并且同时验证移动测量的优势。针对不同的实验搭建相应的实验环境,通过机器人自身CCD相机和天顶全局相机对被测物进行测量,分析测量误差,根据得出的结论去验证本文提出的理论。
[Abstract]:With the development of national defense industry and the improvement of science and technology, intelligent assembly technology of large-scale aerospace products has been developed. Mobile vision measuring robot plays a key role in the intelligent assembly process. Vision measurement and path planning are two important research directions in task planning system of mobile vision measuring robot. For the object to be assembled, we need to determine an optimal observation position of the robot, so that the measurement error of the robot can be minimized, so that the robot can better guide the assembly process. In this paper, a method for determining the optimal observation position of a robot is proposed, which is based on monocular vision. The purpose of path planning is to determine the shortest path of robot moving in the working environment. In this paper, we improve the A * algorithm and propose a path planning method based on the improved A * algorithm. Based on the grid method, three different map construction methods are proposed based on the factors such as the mark points, the measured objects and the observation position in the environment. The main work of this paper is as follows: first, the position planning of mobile vision measuring robot is studied. A monocular visual measurement model was established. According to the principle of monocular vision imaging, the measurement scheme is determined, and based on the geometric relationship and coordinate transformation, the visual measurement models are established for two-dimensional and three-dimensional situations, respectively. An optimal observation position measurement model is established for different working conditions to determine the optimal observation position. Secondly, the path planning method of mobile vision measuring robot is studied. The realization process of A * algorithm is described, and the simulation of path planning method based on A * algorithm in grid map is carried out. To solve the collision problem, the path planning method is improved to avoid the collision problem. Thirdly, the map construction method of mobile vision measuring robot is studied. Different from the traditional map construction, the map construction in this paper introduces the factors such as mark point and object under test, and on this basis, separately considering the factor of mark point, synthetically considering the factor of mark point and object being tested, synthetically considering the mark point, considering separately the factor of mark point, synthetically considering the factor of mark point. There are three different map construction methods: the measured object and the observed position factor. The path planning method based on the improved A * algorithm is used to simulate, and the advantages and disadvantages of the different methods are analyzed and compared. Fourth, the optimal observation position measurement model and path planning method are verified by experiments, and the advantages of mobile measurement are also verified. According to the different experimental environment, the measured object is measured by the robot's own CCD camera and the zenith global camera, the measurement error is analyzed, and the theory proposed in this paper is verified according to the conclusion.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【参考文献】