工业机器人运动学标定及直接示教技术研究

发布时间：2018-06-03 15:52

本文选题：机器人 + 标定　；参考：《安徽工业大学》2017年硕士论文

【摘要】：多自由度串联机械臂广泛应用于工业生产、医疗手术、物流搬运等行业。经过几十年的发展工业机器人技术已经趋于成熟,但随着科技水平的提升,各行业对机器人的性能要求也越来越高。提升机器人性能,成为了当前机器人厂商的迫切目标。本文前半部分采用基于距离误差的运动学标定方法,激光跟踪仪对四轴码垛机器人进行数据测量,辨识出机器人的名义参数误差,补偿后提高了机器人的绝对定位精度。首先使用D-H方法建立待标定机器人的运动学模型,并给出了详细的正、逆运动学推导过程,利用matlab软件进行运动学模型的仿真与验证,验证了运动学模型的有效性。之后分析机器人末端执行器绝对定位误差来源,介绍了距离误差定义,并推导出了由距离误差与连杆参数误差之间的关系,建立距离误差模型。搭建信息采集平台,采集并储数据,依照误差模型建立误差方程。最后用最小二乘法和LM算法对误差方程求解,Levenberg-Marquard算法由于没有二次项舍去误差,相比最小二乘法效果更好,精度提升了35%,距离误差平均值由2.20mm减少到0.27mm。文章后半部分基于力矩控制实现了CY3A冲压机器人的拖动示教。与传统示教方法相比,直接示教操作简单、对工人要求低、示教效率高、能够实现复杂轨迹示教,是示教技术的一个重要研究方向。免力矩传感器的直接示教由于具有明显的成本优势而受到了关注,本文基于力矩平衡的原理来实现机器人的免力矩传感器直接示教。为了补偿示教时的摩擦力矩,建立了关节摩擦模型,使用跟踪微分器获取准确的速度信号,按照库伦—粘滞摩擦模型拟合出了每个关节的摩擦力曲线。对负载较重的关节,利用跟踪微分器得到加速度信号,计算惯性力并进行补偿。补偿后关节所需操作力小于20N,能够实现直接拖动示教。本文对泰禾光电的FR-120A型码垛机器人进行了运动学标定,对CY3A型冲压放料机器人实现了直接示教,论文中的理论也适用于其它类型的工业机器人,相关成果对提高机器人绝对定位精度、实现直接示教等具有指导意义。
[Abstract]:Multi-DOF series manipulator is widely used in industrial production, medical surgery, logistics handling and other industries. After decades of development, the technology of industrial robot has become mature, but with the improvement of science and technology, the performance requirements of robots are becoming more and more high. Improving the performance of robots has become an urgent goal of current robot manufacturers. In the first half of this paper, the kinematics calibration method based on distance error is used to measure the data of the four-axis palletizing robot with laser tracker, and the nominal parameter error of the robot is identified, and the absolute positioning accuracy of the robot is improved after compensation. First, the kinematics model of the robot to be calibrated is established by using D-H method, and the process of forward and inverse kinematics derivation is given in detail. The kinematics model is simulated and verified by matlab software, and the validity of the kinematics model is verified. Then, the source of absolute positioning error of robot end actuator is analyzed, the definition of distance error is introduced, the relationship between distance error and link parameter error is deduced, and the distance error model is established. Set up the information acquisition platform, collect and store the data, and establish the error equation according to the error model. Finally, Levenberg-Marquard algorithm is used to solve the error equation by using the least square method and LM algorithm. Because there are no quadratic terms, the accuracy of Levenberg-Marquard algorithm is better than that of the least square method, and the average distance error is reduced from 2.20mm to 0.27mm. In the second half of this paper, the drag teaching of CY3A stamping robot is realized based on torque control. Compared with the traditional teaching demonstration method, direct teaching operation is simple, the requirement for workers is low, teaching efficiency is high, and it can realize complex trajectory teaching, which is an important research direction of teaching demonstration technology. The direct teaching of torque free sensors has attracted much attention because of its obvious cost advantages. Based on the principle of torque balance, the direct teaching of robot torque free sensors is realized in this paper. In order to compensate the friction moment of teaching, the joint friction model was established, and the accurate velocity signal was obtained by using the tracking differentiator, and the friction curve of each joint was fitted according to the Coulomb viscous friction model. For the joints with heavy load, the acceleration signal is obtained by using the tracking differentiator, and the inertia force is calculated and compensated. After compensation, the required operating force of joint is less than 20 Ns, which can realize direct drag teaching. In this paper, the kinematic calibration of Taihe optoelectronic FR-120A type palletizing robot is carried out, and the direct teaching of CY3A type punching and discharging robot is realized. The theory in this paper is also suitable for other types of industrial robots. The related results are of guiding significance for improving the absolute positioning accuracy of the robot and realizing direct teaching.
【学位授予单位】：安徽工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242.2

【参考文献】