基于嵌入式运动控制器的弧焊机器人控制系统研究

发布时间：2018-03-05 23:03

本文选题：弧焊机器人　切入点：嵌入式运动控制器　出处：《哈尔滨工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：弧焊机器人作为工业机器人中使用最为广泛的一种机器人,其控制系统的研究一直是行业的热点。针对常规示教再现型系统无法满足复杂焊缝加工精度的问题,本文基于GUC-T嵌入式运动控制器设计了一套包括示教再现和视觉传感自动跟踪两大核心功能的六轴弧焊机器人控制系统,借此使用者可以利用示教功能进行非焊缝跟踪任务的示教,利用视觉功能实现焊缝的自动跟踪。围绕该控制系统的设计本课题展开了以下研究工作:首先基于机器人的功能需求和实际工业控制场合特点构建机器人控制系统的硬件结构:针对传统PC系统体积大、实时性差、抗干扰能力弱的问题,本文采用GUC-T嵌入式运动控制器实现机器人的运动规划,同时为了减轻运动控制器的资源负担,引入显控一体计算机构建视觉处理和任务规划子系统模块。对弧焊机器人的运动学模型进行分析。为了满足弧焊机器人焊接过程中的平稳性需求,本文基于雅克比矩阵奇异值分解,提出了一种新的自适应动态阻尼系数构建方式,以此来提高机器人运动规划过程中的奇异鲁棒性,使得机器人在面对奇异问题时能够平稳过渡,在此基础上给出了机器人速度级上的闭环逆解算法。算法通过MATLAB进行仿真验证,并与传统的动态阻尼系数法进行了对比。同样是出于对弧焊机器人运动平稳性的考虑,在轨迹规划时采用正弦曲线优化后的S型加减速算法对空间直线和圆弧进行插补。在姿态规划的过程中,引入单位四元数构造五次样条姿态曲线来实现姿态的高阶连续轨迹规划。针对示教连续轨迹衔接处的速度、加速度突变问题,引入五次贝塞尔曲线来构造过渡段轨迹,并进行了MATLAB的算法仿真验证试验。在视觉跟踪部分,结合点结构光扫描式传感器,构建了手眼方式的焊缝视觉跟踪系统,为了提高手眼标定精度,在五点标定法的基础上利用格拉姆矩阵构造矩阵列向量组的线性无关度度量因子。之后结合分解速度运动控制给出了焊缝实时纠偏的运动控制算法。最后,在硬件平台上搭建嵌入式软件开发环境,为嵌入式运动控制器定制和移植了WinCE操作系统,并基于“HOST/TARGET”嵌入式软件开发模式分别采用QT和VS2008开发嵌入式Linux环境和WinCE环境下的子系统软件,利用TCP/IP协议实现各硬件单元间的信息通讯,通过实验验证了该控制系统的有效性。
[Abstract]:Arc welding robot is one of the most widely used robots in industry. The research of its control system has always been a hot spot in the industry. This paper designs a six-axis arc welding robot control system based on GUC-T embedded motion controller, which includes two core functions: teaching reproduction and visual sensor automatic tracking. With this, the user can use the demonstration function to teach the non-weld tracking task, Based on the design of the control system, the following research work is carried out: firstly, the robot control system is constructed based on the functional requirements of the robot and the characteristics of the actual industrial control situation. Hardware structure: for the traditional PC system volume, In order to reduce the resource burden of motion controller, GUC-T embedded motion controller is used to realize robot motion planning. The visual processing and task planning subsystem module is constructed by introducing the display and control computer. The kinematics model of arc welding robot is analyzed. In order to meet the requirement of stability in the welding process of arc welding robot, Based on the Jacobian matrix singular value decomposition, a new adaptive dynamic damping coefficient construction method is proposed to improve the singular robustness of robot motion planning. On the basis of this, a closed-loop inverse solution algorithm on the velocity level of robot is presented. The algorithm is verified by MATLAB simulation. And compared with the traditional dynamic damping coefficient method, it is also out of the consideration of the stability of the arc welding robot's motion. In the course of trajectory planning, the S-type acceleration and deceleration algorithm after sinusoidal optimization is used to interpolate the spatial straight line and arc. In the process of attitude planning, The cubic spline attitude curve is introduced to realize the high-order continuous trajectory planning of the attitude, and the fifth Bessel curve is introduced to construct the transition trajectory for the problem of the velocity and acceleration abrupt change at the junction of the teaching continuous trajectory. In the part of visual tracking, combining point structured light scanning sensor, the visual tracking system of weld seam in hand-eye mode is constructed, in order to improve the precision of hand-eye calibration, the algorithm of MATLAB is simulated and verified. On the basis of the five-point calibration method, Gramm matrix is used to construct the linear independence measure factor of matrix vector group. Then, a real-time motion control algorithm for weld seam correction is given in combination with decomposed velocity motion control. The embedded software development environment is built on the hardware platform, and the WinCE operating system is customized and transplanted for the embedded motion controller. Based on "HOST/TARGET" embedded software development mode, QT and VS2008 are used to develop subsystem software in embedded Linux environment and WinCE environment, and the information communication between hardware units is realized by TCP/IP protocol. The effectiveness of the control system is verified by experiments.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】