作　　者： (邹风山）; (赵彬）;

机构地区： 沈阳新松机器人自动化股份有限公司,沈阳110168 中国科学院沈阳自动化研究所,沈阳110016

出　　处： 《组合机床与自动化加工技术》 2017年第8期100-104,共5页

摘　　要： 文章解决了电机电流进行机器人主动柔顺控制的问题,同时对七自由度关节一体机器人进行拖曳式示教实验。首先针对关节一体化机器人进行运动学分析,采用多层前向神经网络解决机器人运动学冗余轴反解问题,将复杂的非线性优化问题转化为简单的代数方程组求解问题。然后由于机器人特殊的构型及关节耦合复杂,采用牛顿—欧拉动力学建模。将永磁同步电机的转矩模型与机器人动力学模型结合,得到基于电流模型的力检测。最后实现基于电机电流的机器人主动柔顺控制,并在七自由度关节一体机器人上进行拖曳式示教实验,实际运行结果验证了算法的有效性。 Based on the motor current, the active compliant control of the robot is studied, and the drag- and-drop teaching experiment is carded out for the 7-DOF joint one-body robot. Firstly, the multi-layer feedforward neural network is used to solve the inverse kinematics problem of the robot kinematics, and the complex nonlinear optimization problem is transformed into a simple algebraic equation group for solving the kinematic analysis of the joint-integrated robot. Then, due to the special configuration of the robot and the complicated coupling of the joints, the Newton-Euler dynamics model is adopted. The torque model of the permanent magnet synchronous motor is combined with the robot kinetic model to obtain the force detection based on the current model. Finally, the active compliant control of the robot based on the motor current is realized, and the teaching experiment of the robot is carded out. The actual running result verifies the effec- tiveness of the method.

关 键 词： 电流力控制 神经网络 机器人动力学 拖曳示教