基于真实工况下的工业机器人关节转角辨识与分析

本文基于多目视觉测量系统，对真实工况下连续运动的工业机器人进行关节转角的实时重构．该方法通过机器人运动前后的坐标集，在对刚体运动进行最优拟合的条件下，采用最小二乘法获得了各关节的旋转矩阵与平移向量．在此基础上，在考虑相邻关节牵连运动的前提下，获得了各关节的相对旋转矩阵．结合罗德里格斯变换理论通过相对旋转矩阵，确定了各关节转角．仿真与实验分析，验证了该方法的有效性与正确性．在该测量与辨识体系下，初步确定了各关节转角随机器人运动的真实状态．2与3杆臂由于物理尺寸呈细长形状，连杆挠度较大，这时变形误差与振动建立了关系，角振动幅度很大，曲线随机性较强．其余杆臂由于刚度较大，关节转角曲线呈光滑状态．

Joint Rotation Angle Identification and Analysis for Industrial Robots Based on Real Working Conditions

Based on the multi-vision measurement system, the real-time reconstruction of the joint rotation angle of the industrial robots in real working conditions is carried out. In this approach, the rotation matrix and translation vector of each joint are obtained by the least square method under the condition of optimal fitting of rigid body motion through the coordinate set before and after the robot motion. On this basis, the relative rotation matrix of each joint is derived considering the implicative motion of adjacent joints. Combined with Rodriguez's transformation theory, the angle of each joint is deduced by the relative rotation matrix. Simulation and experimental analysis validate the effectiveness and correctness of the proposed method. Based on the measurement and identification system, the real state of each joint angle is determined. Due to the slender physical shape of the 2 and 3 arms, the deflection of the connecting rod is large. At this time, the relationship between deformation error and vibration is established. The angular vibration range is large, and the randomness of the curve is significant. Due to the larger stiffness of the other arms, the joint angle curve is smooth.