科技的进步，机器人行业不断发展。在工业领域，工业机械臂能够进行点焊、弧焊、装配、喷漆、切割、搬运等各种精度操作；在航空航天领域，空间机械臂用于代替宇航员在太空进行科学试验、出舱操作、维修设备故障以及空间探测等活动。为了保障机械臂的工作效率与工作稳定性，需要对机械臂进行标定以及运动学参数辨识与补偿，尽可能减小机械臂的误差，使机械臂完成高精度操作。本课题以六自由度机械臂展开研究，对其运动学参数进行了辨识；然后根据自由度由少到多，参数由简到繁，依托国家863某重点机械臂项目，以空间多自由度机械臂为研究对象，对该机械臂系统进行了运动学测量标定，完成了机械臂系统的辨识。主要研究内容如下：首先，基于MD-H法建立六自由度机械臂运动学模型，推导出末端位姿误差与机械臂关节参数误差之间的微分方程，对六自由度机械臂进行了误差建模。分析微分误差情况，由泰勒级数原理，推导得出了连杆型机械臂末端位姿误差与连杆参数之间的一阶微分和二阶微分误差的一般形式，为后续多自由度机械臂误差建模和参数转换提供了理论基础。针对六自由度机械臂进行仿真和实验，在仿真层面，通过Qt开发环境编写了一套六自由度机械臂的软件控制平台，验证了六自由度机械臂正逆解模型的准确性，通过此平台获取机械臂的理论位置数据。基于六自由度UR5机器人与API激光跟踪仪标定实验平台，设计实验，改进了参数测量方法，通过最小二乘法，辨识出了模型的误差。由六自由度推广到12自由度机械臂，首先完成了机械臂硬件平台的搭建，利用Optitrack测量系统，对机械臂进行了原始参数测量。通过对多自由度机械臂求解逆解，将关节空间参数转换为驱动空间参数，简化了待辨识的参数，进而建立了数据集。基于反向传播网络的基本原理，确定了机械臂参数辨识的神经网络模型，完成该系统上位机开发和参数辨识实验，并对实验结果进行了验证分析。
The advancement of technology and the continuous development of the robot industry. In the industrial field, industrial robotic arms can perform various precision operations such as spot welding, arc welding, assembly, painting, cutting, and handling. In the aerospace field, space robotic arms are used to replace astronauts in space for scientific experiments and outbound operations. , maintenance equipment failures and space exploration activities. In order to ensure the working efficiency and working stability of the manipulator, it is necessary to calibrate the manipulator and identify and compensate the kinematic parameters, minimize the error of the manipulator, and complete the high-precision operation of the manipulator.This subject is studied with a six-degree-of-freedom mechanical arm, and its kinematic parameters are identified. Then, according to the degree of freedom, the parameters are from simple to complex, relying on the national key mechanical arm project of 863, and the space multi-degree of freedom mechanical The arm is the research object, and the kinematic measurement calibration of the robot arm system is completed, and the identification of the robot arm system is completed. The main research contents are as follows:Firstly, the kinematics model of the six-degree-of-freedom manipulator is established based on the MD-H method, and the differential equation between the end pose error and the mechanical arm joint parameter error is derived. The error of the six-degree-of-freedom manipulator is modeled. By analyzing the differential error condition, the general form of the first-order differential and second-order differential error between the end pose error of the link type manipulator and the link parameters is derived from the Taylor series principle, which is the subsequent multi-degree-of-freedom manipulator. Error modeling and parameter conversion provide a theoretical basis.The simulation and experiment of the six-degree-of-freedom manipulator are carried out. At the simulation level, a software control platform of a six-degree-of-freedom manipulator is programmed through the Qt development environment to verify the accuracy of the positive and negative solution model of the six-degree-of-freedom manipulator. The platform acquires theoretical position data of the robot arm. Based on the calibration platform of the six-degree-of-freedom UR5 robot and API laser tracker, the design experiment was carried out, and the parameter measurement method was improved. The error of the model was identified by the least squares method.From the six degrees of freedom to the 12-DOF manipulator, the mechanical arm hardware platform was first built, and the original parameters of the manipulator were measured using the Optitrack measurement system. By solving the inverse solution to the multi-degree-of-freedom manipulator, the joint space parameters are transformed into the driving space parameters, which simplifies the parameters to be identified, and then establishes the data set. Based on the basic principle of BP, the neural network model of the parameter identification of the manipulator is determined. The development of the host computer and the parameter identification experiment are completed, and the experimental results are verified.
