五自由度搬运机器人的运动控制与精度标定方法研究
文献类型：学位论文
| 作者 | 丁磊
|
| 学位类别 | 工学硕士 |
| 答辩日期 | 2017-05-19 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 北京 |
| 导师 | 谭民 ; 李恩 |
| 关键词 | 搬运机器人 运动学分析 控制系统 ROS 运动学标定 |
| 中文摘要 | 目前在冲压生产行业中，采用搬运机器人代替人工操作的自动化生产方式是提升产品质量、保障安全生产、提高生产效率、降低生产成本的技术解决途径，也是该行业制造模式发展的主要趋势。但由于冲压行业对生产成本的敏感性，传统工业机器人在成本和性能方面很难达到良好的平衡，而低成本搬运机器人的工作空间和定位精度等方面又很难满足自动化生产的需求。因此，研究面向低成本冲压搬运机器人的优化设计和高精度控制问题具有重要意义。 本文以五自由度搬运机器人为研究对象，主要开展控制系统设计、运动学分析、轨迹规划、精度参数标定等四个方面的研究工作，形成具有自主知识产权的冲压机器人控制系统，提出适用于冲压搬运过程的轨迹规划方法，结合基于立体视觉的运动学参数标定方法，提高机器人的定位精度，实现低成本机器人在冲压搬运过程中的连续运动功能和绝对定位性能。本文的主要内容有： (1) 结合冲压生产线实际情况，设计满足冲压搬运作业要求的机器人机械结构，并进一步开展搬运机器人的正逆运动学建模，工作空间分析和精度误差分析。采用D-H参数法对机器人的运动学进行建模，分析了逆运动学的求解问题，并通过基于MATLAB机器人工具箱的仿真，验证了运动学模型的正确性；基于运动学模型和各关节的运动范围，得到了搬运机器人工作空间的显示表达式；通过建立机器人雅可比矩阵，得到了末端定位精度与各关节误差的关系，为加工、装配和控制的精度提供了参数指导。 (2) 在控制系统整体需求分析的基础上，基于嵌入式系统自主设计了满足冲压搬运作业要求的低成本机器人控制系统，并提出了可实现多路精确控制的多轴运动控制器设计方法。控制系统的三个核心模块：主控制器、运动控制卡和手动脉冲发生器的软硬件及通信协议均为自主开发。主控制器的人机交互界面利用按钮式操作方式实现，不使用任何编程语言，仅通过点击和选择就可完成示教过程，操作简单，易于使用；完全自主开发的运动控制卡充分利用ARM处理器内部定时器资源，自主设计脉冲输出控制电路，通过定时器的定时、脉冲计数、脉冲输出等功能，实现多路脉冲数的精确控制和输出，进而实现机器人的精确运动控制；手动脉冲发生器使手动控制脱离人机界面，提高操作的简便程度，确保现场操作的安全性。 (3) 针对冲压搬运过程的运动路径要求，对轨迹规划方法进行研究，总结搬运机器人基本工序，将笛卡尔空间轨迹规划和关节空间轨迹规划方法相结合，提出了适用于冲压搬运过程的轨迹规划方法。在搬运机器人机构设计及运动学分析、控制系统设计和轨迹规划方法研究基础上，完成了控制系统和规划方法的集成和测试，基于ROS(Robot Operating System)搭建了机器人的半实物仿真系统，通过将各关节运动数据与仿真模型相结合，利用实际控制系统采集到的各关节运动数据驱动ROS中机器人模型进行三维运动，验证了控制系统的有效性。 (4) 为提高搬运机器人的末端定位控制精度，提出了一种低成本的基于最优轨迹和单目视觉的运动学标定方法，对系统装配后的参数误差进行标定和补偿。针对自主设计的由旋转关节和平移关节组成的搬运机器人结构，利用MDH法对运动学模型进行修正，得到机器人的误差模型；通过一种新型单目视觉测量系统，利用单目摄像机配合安装在机器人末端的靶标实现末端位置测量，有效保证测量精度的同时大幅降低了测量系统的成本，然后利用遗传算法在工作空间中得到用于运动学标定的搬运机器人最优轨迹，实现在最少测量点的前提下保证标定结果最优。在此基础上，利用Levenberg-Marquard算法标定机器人的结构参数，使机器人绝对定位精度达到要求。 |
| 英文摘要 | At present, in the stamping production industry, automatic production mode which uses handling robots instead of manual operations is the technical solution to promote product quality, guarantee safety in production, improve production efficiency and reduce production cost. It is also the main trend of the production mode in this industry. But due to the sensitivity of production cost in stamping industry, it is difficult for the traditional industrial robot to achieve good balance in terms of cost and performance. Besides, it is also difficult to meet the requirements of automatic production for low cost handling robot because of its workspace and positioning precision. Therefore, the research of the optimal design and high precision control of the low cost handling robot is of great significance. Based on the five degree of freedom handling robot, the research mainly focuses on the control system design, kinematics analysis, trajectory planning and the kinematic calibration method. The objectives of this paper are accomplishing the control system for handling robot with independently intellectual property rights, proposing trajectory planning algorithm which can be applied to stamping process, improving the absolute positioning precision of handling robot by using the kinematic calibration method based on stereo vision system, and realizing the continuous motion control and absolute positioning of the low cost handling robot in the stamping process. The main contents of this paper are: (1) According to the actual situation of punching production line, handling robot mechanical structure was designed to meet the requirements of the punching production. Kinematics modeling, workspace and error analysis were further developed. The forward kinematic model was established by using the D-H method. The inverse kinematic was solved. And then the simulation of handling robot was performed based on MATLAB and the robotics toolbox to verify the correctness of the kinematic model; Based on the kinematics model and the motion range of each joint, the explicit expression of the workspace of handling robot was obtained; By establishing the Jacobian matrix of handling robot, the relationship between the positioning accuracy and the error of each joint was obtained. This provided instructive parameters for mechanical process, assembly and robot control. (2) According to the overall requirements analysis of robot control system, the low cost handling robot control system was designed to meet the requirements of the punching production based on the embedded system. And the design method of multiaxis motion controller was put forward to realize the precision control of multiple motors. The three core modules of the control system were the main controller, the motion control card and the manual pulse generator. The human-machine interface of the main controller was realized by the operation mode of buttons instead of programming languages. The teaching process was completed only by click and selection. The advantages are simple to operate and easy to use; the independently developed motion control card made full use of the internal timers in the ARM processor. And the pulse output control circuit was independently designed. Through the timing, pulse counting and pulse output function of the timer, the accurate control and output for pulses were realized, and then the accuracy motion control of handling robot was realized; Manual pulse generator disengaged manual control from the human-machine interface to improve the convenience of operation and ensure the safety of on-site operation. (3) According to the requirements of stamping process on the motion path of handling robot, research on trajectory planning algorithm was done. Trajectory planning algorithm which could be applied to handling in the stamping process was proposed by summarizing the basic handling process and combining the trajectory planning algorithm in the Descartes space and the joint space. Based on the mechanical design, kinematics analysis, control system design and trajectory planning algorithm, the integration and test of the control system were completed. The hardware-in-the-loop simulation system for handling robot was built based on ROS (Robot Operating System). By combining the simulation model and movement data of each joint, the joint data which were collected in the actual control system were used to drive the simulation model in ROS to realize 3D motion. The effectiveness of the control system was verified. (4) In order to improve the absolute positioning precision of handling robot, a low cost novel kinematic calibration method based on optimal trajectory planning and monocular vision was proposed to perform calibration and compensation for the parameters error after assembly. According to the independently developed mechanical structure composed of three rotary joints and two translational joints, the forward kinematic model was modified by the MDH method to establish the error model of the handling robot; a low cost monocular vision measurement system was designed with monocular camera and self-designed target to realize position measurement. The measurement system can effectively guarantee the accuracy of position measurement and greatly reduces the cost. And then the optimal trajectory for kinematic calibration was obtained by Genetic Algorithm in the workspace of handling robot to realize the optimal calibration results under the premise of the minimum measurement points. On this basis, the Levenberg-Marquard algorithm was used to calibrate the structural parameters of the handling robot to meet the requirement of absolute positioning precision. |
| 源URL | [http://ir.ia.ac.cn/handle/173211/14630]  |
| 专题 | 毕业生_硕士学位论文 |
| 作者单位 | 1.中国科学院自动化研究所 2.中国科学院大学 |
| 推荐引用方式 GB/T 7714 | 丁磊. 五自由度搬运机器人的运动控制与精度标定方法研究[D]. 北京. 中国科学院研究生院. 2017. |
入库方式： OAI收割
来源：自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。