孙定阳, 沈浩, 郭朝, 肖晓晖. 绳驱动柔性上肢外骨骼机器人设计与控制[J]. 机器人, 2019, 41(6): 834-841. DOI: 10.13973/j.cnki.robot.180750
引用本文: 孙定阳, 沈浩, 郭朝, 肖晓晖. 绳驱动柔性上肢外骨骼机器人设计与控制[J]. 机器人, 2019, 41(6): 834-841. DOI: 10.13973/j.cnki.robot.180750
SUN Dingyang, SHEN Hao, GUO Zhao, XIAO Xiaohui. Design and Control of the Cable Driven Compliant Upper Limb Exoskeleton Robot[J]. ROBOT, 2019, 41(6): 834-841. DOI: 10.13973/j.cnki.robot.180750
Citation: SUN Dingyang, SHEN Hao, GUO Zhao, XIAO Xiaohui. Design and Control of the Cable Driven Compliant Upper Limb Exoskeleton Robot[J]. ROBOT, 2019, 41(6): 834-841. DOI: 10.13973/j.cnki.robot.180750

绳驱动柔性上肢外骨骼机器人设计与控制

Design and Control of the Cable Driven Compliant Upper Limb Exoskeleton Robot

  • 摘要: 为了提高上肢外骨骼机器人的拟人化程度及关节柔性,设计了一种由串联弹性驱动器和鲍登线驱动的4自由度柔性上肢外骨骼机器人.首先,设计一种六连杆双平行四边形机构,建立肩关节虚拟转动中心,满足人体肩部3自由度运动需求.然后,设计基于串联弹性驱动器和鲍登线的驱动模块,将驱动器和机器人关节分离,降低结构的复杂度,减轻关节质量,实现力矩/位置信息的反馈.最后,构建机器人运动学及动力学模型,设计关节阻抗控制器并对样机肘关节进行阻抗控制实验.由实验结果可知,刚度系数在0.5 N· m/(°)~1.5 N· m/(°)时,力矩跟踪均方根为0.33 N· m;阻尼系数在0.001 N· m· s/(°)~0.01 N· m· s/(°)时,力矩跟踪均方根为0.57 N· m.实验结果表明,调节阻抗控制器中的阻抗系数能够改变关节的刚度和阻尼特性,从而提高人机连接的柔顺性.因此该机器人可以满足康复训练需求.

     

    Abstract: In order to improve the anthropomorphosis level and flexibility of the upper limb exoskeleton robot, a compliant 4-DOF (degree of freedom) upper limb exoskeleton robot driven by serial elastic actuator (SEA) and Bowden cable is designed. Firstly, A six-link double parallelogram mechanism is designed to establish a virtual rotational center and realize 3-DOF motion of human shoulder. Then, a drive module based on the SEA and Bowden cable is designed to separate the actuator and the robot joint, reduce the complexity of the structure, relieve the mass of joints and get the feedback of the torque and position information. Finally, the kinematics and dynamics model of the robot is built for developing an impedance controller. The impedance control test is carried out on the elbow joint. The results show that the root mean squared error of the torque tracking is 0.33 N·m when the virtual stiffness coefficient is in the range of 0.5 N·m/(°)~1.5 N·m/(°), the root mean squared error of the torque tracking is 0.57 N·m when the virtual damping coefficient is in the range of 0.001 N·m·s/(°)~0.01 N·m·s/(°). The result shows that the stiffness and damping of the elbow joint can be changed by adjusting the impedance coefficient of the impedance controller, which promotes the compliance of the human-robot interaction. So the robot can meet the needs of rehabilitation training.

     

/

返回文章
返回