机械臂运动学参数的高精度标定方法研究

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 机器人学研究室

题名	机械臂运动学参数的高精度标定方法研究
作者	刘辉
学位类别	硕士
答辩日期	2016-05-25
授予单位	中国科学院沈阳自动化研究所
导师	李斌
关键词	机械臂运动学参数减速比标定
其他题名	Research on the High precision method of calibration of the kinematic parameters of manipulator
学位专业	控制工程
中文摘要	本文依托于国家科技支撑计划“国际热核聚变实验堆（ITER）计划专项”的子项目“核聚变主从精细遥操作机器人”，对机械臂末端绝对定位部分开展研究。本文旨在通过相关研究，提出一种能够较高精度地辨识并且补偿机械臂运动学参数误差的方法，为实现机械臂精确定位提供支撑。针对于此，本文的主要研究内容如下：首先对国内外目前关于机械臂运动学参数标定方法的研究作了系统的研究分析。标定过程主要分为模型建立、位姿测量、参数辨识、误差补偿等四个步骤，本文的相关研究也主要围绕着四个步骤展开。机械臂的运动学模型有D-H模型及其改型、连杆模型、旋量模型等，后者主要考虑的是平行关节、耦合关节等特殊情况。本文中机械臂为串联旋转关节结构，因此采用适用于绝大多数机械臂的D-H模型进行机械臂运动学建模，这样做也可提高研究成果的可移植性和通用性。位姿测量主要分为开环法、闭环法和旋转轴测量法等。开环法使用了机械臂之外的仪器设备对末端位姿进行测量，如三维坐标采集和位置跟踪系统等，测量准确度高且通用性强，因此本文最终采用三维运动测量仪搭建测量系统，对机械臂末端位姿进行测量。学界对参数辨识的研究较多，如最小二乘和迭代、基于距离的辨识、两步法、递推辨识、智能辨识以及几何法辨识等等，各有利弊，需要针对机械臂的具体情况选择相应的算法。另外，从辨识算法的发展轨迹也可以看出，现有的辨识算法还不能称得上完美，而是在不断地改善进步的。针对多自由度串联机械臂的情况，采用几何辨识策略有利于逐步控制辨识误差，且更具有通用性。机械臂运动学参数误差补偿主要分为关节空间补偿、参数空间补偿及基于智能算法的补偿等几类。本文基于解超定微分方程组的方法，利用机械臂运动学模型参数的名义值和标定值，通过计算关节补偿角，从而在关节空间对参数误差进行补偿，避免了雅各比矩阵的求解，提高了计算效率。根据以上研究，设计实现了一套机械臂运动学参数标定系统，为所提标定方法的实验及应用提供了支撑。其中包括了基于红外三维运动测量仪和HP xw8200工作站搭建硬件系统，和基于MATLAB语言，采用模块化、图形化设计实现的软件系统。最后，采用本文提出的几何标定方法，对核聚变主从精细遥操作机器人手臂部分的运动学参数进行了标定实验。在实验的过程中发现关节电机减速比的误差也会在一定程度上影响机械臂的绝对定位精度，并对其进行了标定。在误差补偿实验中，综合考虑机械臂的运动学模型参数误差和关节电机减速比误差，设计了对比实验，实验结果较为理想，验证了标定方法的有效性。
英文摘要	This paper relies on the“Nuclear fusion subordination precise tele-operation Robot”, a subproject of the national science and technology support program "The International Thermonuclear Experimental Reactor (ITER) plan", and research on the part of the manipulator end absolute positioning. The purpose of this paper is to through relevant study, design a system to highly precisely udentification and compensation of the errors of the kinematic parameters of the manipulator, to lay the foundation for the realization of the precise positioning of the manipulator. In view of this, the main research contents of this paper are as follows: First of all, systematically research and analysis the work on manipulator kinematic parameter calibration methods at home and abroad. The calibration process is mainly divided into four steps: model establishment, position and attitude measurement, parameter identification and error compensation. Related research in this paper is mainly around this four steps. Manipulator kinematic model has D-H model and its improved type, Connected rod model, and Rotary model, etc. The latters are mainly considered parallel joints, coupling joints and other special circumstances. The joints of the manipulator in this paper is connected in series, so the D-H model which is suitable for the vast majority of the manipulator is used to carry out the kinematics modeling of the manipulator in this paper. In doing so, also improves the portability of the research results of the calibration system. Pose measurement is divided into open loop measurement, close measurement and the rotation axis measurement, etc. Open loop measurement use equipments outside the manipulator, such as 3D coordinate acquisition and position tracking system, etc, accuracy is higher and the adaptability is better, so in this paper, we use the three-dimensional motion measurement system to build the measurement system, to measure the end joint position of the manipulator. The research of parameter identification is more in the academic field, such as the least squares and iteration, identification based on distance, two steps identification method, recursive identification, intelligent identification and geometric identification and so on, each has advantages and disadvantages, need to select the appropriate algorithm for the specific circumstances of the manipulator. In addition, through the trajectory of the development of the identification algorithm, we can see that the existing identification algorithm can not be called perfect, but in the continuous improvement of the progress. According to the situation of the multi degree of freedom series manipulator, the geometric identification method is with a higher precision, and is more universal. The error compensation of the kinematic parameters of the manipulator is mainly divided into joint space compensation, parameter space compensation, and compensation based on Intelligent Algorithm,etc. In this paper, based on the method of solving the set of super fixed differential equations, with the nominal value and the calibration value of the kinematic model of the manipulator, by calculating the joint compensation angle, in order to compensate the error of parameters in joint space, and the solution of Jacoby matrix is avoided and the computational efficiency is improved. According to the research, a system of kinematic parameters calibration system is designed, laying the foundation for the experiment and application of the calibration method. Among it a hardware system based on infrared 3D motion measurement instrument and xw8200 HP workstation is biuld, and a software system based on MATLAB language, modular, and graphical design, is implemented. At last, use the proposed geometric calibration method calibrated the manipulator part of the Nuclear fusion subordination Precise Remote operat Robot. In the course of the experiment, realise that the error of the reduction ratio of the joint motor can also affect the absolute positioning accuracy of the robot arm to a certain extent, and carry out its calibration. In the experiment of error compensation, the kinematic model parameter error of the manipulator and the reduction ratio error of joint motor are considered, and the comparative experiment was designed. The experimental results are satisfactory, which verifies the validity of the calibration scheme.
语种	中文
产权排序	1
页码	69页
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/19670]
专题	沈阳自动化研究所_机器人学研究室
推荐引用方式 GB/T 7714	刘辉. 机械臂运动学参数的高精度标定方法研究[D]. 中国科学院沈阳自动化研究所. 2016.