6自由度关节机器人运动学反解的共形几何代数方法

发布时间：2018-01-06 11:24

本文关键词：6自由度关节机器人运动学反解的共形几何代数方法　出处：《机械科学与技术》2017年08期 　论文类型：期刊论文

【摘要】：机器人运动学模型的建立普遍利用Denavit-Hartenberg(D-H)参数法和旋量法,但是D-H参数法的几何意义不够明确,旋量法则有低自由度、关节轴线相交于1点的结构限制。针对6自由度关节机器人反解问题提出一种利用共形几何代数求解的新方法。首先以旋转关节轴线和旋转平面为基础建立无坐标系机器人模型,并定义肩部、肘部和腕部3种机器人结构设计形式。然后在上述模型和机器人结构下利用已知共形点建立直线、平面、圆周和球体等共形几何对象,通过几何对象的约束关系进行简单的代数运算完成各关节轴的旋转角计算。另外,该方法在进一步简化代数运算的基础上,利用2直线对象和旋转平面法矢量的约束关系唯一确定旋转角,从而完成运动学反解计算。最后,以常用的人机协作的UNIVERSAL ROBOT UR5 6自由度关节机器人为例,利用该算法进行运动学反解的验证,计算结果表明该算法的正确性。
[Abstract]:Denavit-Hartenberg-D-H parameter method and spinor method are widely used to build kinematics model of robot, but the geometric significance of D-H parameter method is not clear enough. The spinor law has a low degree of freedom. This paper presents a new method for solving the inverse solution of a 6-DOF joint robot by using conformal geometric algebra. Firstly, based on the axis of rotating joint and the plane of rotation, a sit-free method is established. Standard robot model. Three kinds of robot structures are defined, such as shoulder, elbow and wrist. Then the conformal geometric objects such as line, plane, circumference and sphere are established by using the known conformal points under the above model and robot structure. The rotation angle of each joint axis is calculated by the simple algebraic operation of the constraint relation of the geometric object. In addition, the method is based on further simplifying the algebraic operation. The rotation angle is uniquely determined by the constraint relation between the 2-line object and the normal vector of the plane of rotation, and the inverse kinematics solution is calculated. Finally. Taking the commonly used UNIVERSAL ROBOT UR5 6 DOF joint robot as an example, the algorithm is used to verify the kinematics inverse solution. The results show that the algorithm is correct.
【作者单位】：常州工学院机械与车辆工程学院;
【基金】：江苏省自然科学基金青年项目(BK20140252) 江苏省高校自然科学研究面上项目(14KJB510003)资助
【分类号】：TP242
【正文快照】： 机器人的反解[1-3]是在已知机器人末端执行器即工具的期望位姿的前提下,求解机器人各关节的旋转角的问题。它是进行机器人轨迹规划和实时控制的前提和基础,许多学者在这方面做了大量的工作[4-8]。机器人反解首先需要建立机器人运动学模型,目前机器人运动学分析方法主要有2种:D

【相似文献】