基于机器视觉的Tripod机器人控制研究

发布时间：2018-03-08 23:00

本文选题：Tripod并联机器人　切入点：机器视觉　出处：《东华大学》2017年硕士论文　论文类型：学位论文

【摘要】：目前传统的工业分拣过程存在着高成本和低效率的缺点,无法达到理想的效果,随着机器人在工业领域的应用,不仅可以提高工业过程的生产效率,还能降低劳动力成本。工业机器人根据其结构主要可以分为串联机器人和并联机器人两种,并联机器人由于其速度快、精度高、承载能力强的优点,被广泛应用于大质量负载、精密定位和高速运动等场合,特别是一些工业分拣的应用。本文以贝加莱(BR)公司设计的Tripod并联机器人为对象,主要研究了三自由度机器人的视觉识别、轨迹规划和运动控制等问题,具体内容如下:(1)设计了基于LabVIEW的Tripod机器人视觉系统。针对Tripod机器人的系统结构和操作平台,结合机器视觉和图像处理,对平台几何体的颜色和形状进行识别,应用图像边缘提取、滤波平滑、中心点检测等算法,利用LabVIEW强大的视觉模块和丰富的库函数,对图像进行了相应的预处理、目标特征提取以及中心点定位等工作。(2)建立了Tripod并联机器人的运动学和动力学模型。对机器人的几何模型分别从运动学和动力学两个方面进行了解析,针对正向运动学、反向运动学和雅可比矩阵对三自由度并联机器人进行了过程分析和推理;在建立的Tripod机器人运动学模型基础上,结合力的相互作用关系建立了机器人的动力学模型。(3)进行了Tripod机器人工作空间分析和末端轨迹规划。根据推导得出的Tripod机器人正向运动学公式,利用旋转坐标系确定了机器人末端执行器可达临界空间,并据此对机器人的工作空间进行了仿真分析。对机器人末端的轨迹进行了规划,在基本的直线和圆弧插补的基础上提出了连续椭圆轨迹规划,并利用五次多项式构造出了末端执行器轨迹函数,对Tripod机器人单位化的轨迹进行了仿真。(4)利用基于李雅普洛夫稳定的自适应控制算法对Tripod机器人进行了位置跟踪。基于建立的机器人动力学模型,在MATLAB中利用S函数搭建了Tripod机器人的控制模型,在给定的输入位置情况下,针对机器人自适应控制算法进行了仿真,在基于全局稳定的自适应控制基础上设计了减小位置误差的自适应控制器,结果显示最终能够准确跟踪关节的位置和速度。
[Abstract]:At present, the traditional industrial sorting process has the shortcomings of high cost and low efficiency, so it can not achieve the ideal effect. With the application of robot in the industrial field, it can not only improve the production efficiency of industrial process, Industrial robot can be divided into series robot and parallel robot according to its structure. Parallel robot is widely used in mass load because of its advantages of high speed, high precision and strong bearing capacity. Precision positioning and high speed motion, especially the applications of industrial sorting, are studied in this paper. In this paper, the vision recognition of 3-DOF robot is studied, which is based on the Tripod parallel robot designed by Begaray Brr Company. The main contents of this paper are as follows: 1) the Tripod robot vision system based on LabVIEW is designed. Aiming at the system structure and operation platform of Tripod robot, combined with machine vision and image processing, The color and shape of the platform geometry are recognized, the image edge extraction, filter smoothing, center point detection and so on are used. The image is preprocessed by the powerful visual module of LabVIEW and the rich library function. The kinematics and dynamics models of Tripod parallel manipulators are established. The geometric models of the robot are analyzed from two aspects of kinematics and dynamics respectively. The inverse kinematics and Jacobian matrix are used to analyze and infer the process of 3-DOF parallel robot. Based on the kinematics model of Tripod robot, The dynamic model of Tripod robot is established by the interaction relation of binding force. The workspace analysis and terminal trajectory planning of Tripod robot are carried out. According to the forward kinematics formula of Tripod robot, the forward kinematics formula of Tripod robot is derived. The reachable critical space of the robot end actuator is determined by using the rotating coordinate system, and the workspace of the robot is simulated and analyzed accordingly, and the trajectory of the robot end is planned. Based on the basic interpolation of straight lines and arcs, the continuous elliptic trajectory planning is proposed, and the terminal actuator locus function is constructed by using the quintic polynomial. This paper simulates the unit trajectory of Tripod robot. It uses the adaptive control algorithm based on Lyapunov stability to track the position of Tripod robot. In MATLAB, the control model of Tripod robot is built by using S-function. The adaptive control algorithm of the robot is simulated under the given input position. An adaptive controller to reduce the position error is designed on the basis of the adaptive control based on global stability. The results show that the position and speed of the joint can be tracked accurately.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】