中国科学院自动化研究所机构知识库

近年来，协作机器人的设计、规划与控制技术快速发展。为了增强机器人的协同作业能力，需要在传统的运动控制和动力学分析基础上，提高机器人运动规划的实时性和光滑性，改善牵引控制效果，并增强其安全性能。本文针对多关节机器人动态位姿规划、动力学参数校正、无力/力矩传感器的牵引示教控制和人机协作型机械臂的设计与控制等方面展开研究。主要内容如下：

一、针对多关节机器人的动态位姿规划问题，提出了相邻运动轨迹之间的过渡方法，包括位置过渡和姿态过渡。位置过渡方法基于贝塞尔曲线实现了直线和圆弧之间的光滑过渡，保证了G2连续性。姿态过渡方法中，给出了球面贝塞尔控制点的选取方法，证明了轨迹的光滑性，并采用迭代计算进行插补，使得旋转轴和角速度光滑过渡到目标状态。实验表明，该方法能够实时平滑机器人轨迹，减少停顿。

二、针对动力学参数校正问题，提出了参数校正方法的评价指标，比较了常用的惯性参数和摩擦力参数校正模型，分析了不同方法的优缺点。通过仿真和实验，比较了不同参数计算方法的结果，并以六自由度工业机器人ER20-C10为例说明了模型和计算方法的选择策略。

三、针对无力/力矩传感器的牵引示教控制问题，提出了多关节机器人的外力估计和外力跟随控制方法。分析了驱动系统和机械臂连杆的惯性力和摩擦力，建立了非线性模型，并使用最优化方法校正了参数。牵引示教中，提出了基于虚拟质量和虚拟摩擦力的牵引控制方法，使得末端执行器能够按照外力的方向运动并抑制抖动。实验结果表明，使用校正后的参数能够准确地计算外力，达到较好的牵引控制效果。

四、针对人机协作型机械臂的设计与控制问题，设计了机器人的结构和控制系统，减小了连杆的惯性，分析了在关节扭矩约束范围内实现路径跟随的方法和基于反馈线性化方法的柔性关节轨迹跟踪控制，提高了人机协作的安全性，减小了跟踪误差。最后，将所提轨迹规划和动力学控制方法集成在该机器人系统上进行了多项实验，验证了本文所提方法和设计的有效性。

最后，总结了本文所取得的成果，并对下一步需要开展的研究做了展望。

In recent years, the design, planning, and control technology of collaborative robots have developed rapidly. To enhance the cooperation ability of the robot, it is necessary to improve the real-time performance and smoothness of the motion planning based on the traditional motion control and dynamics analysis, achieve better guiding performance, and enhance the safety. This thesis focuses on the dynamic pose planning, dynamic parameter calibration, sensorless guiding control, and the design and control method of collaborative manipulator. The main contents are as follows:

Firstly, a transition method between adjacent trajectories is proposed to deal with the problem of dynamic pose planning of multi-joint robots, including position transition and orientation transition. Based on the Bezier curve, the position transition method achieve smooth transitions between lines and circles, ensuring G2 continuity. In the orientation transition, the control point selection method of the spherical Bezier curve is given and the smoothness of the trajectory is proved. An iterative calculation method is used to interpolate. The rotation axis and angular velocity transit to the target state smoothly. Experiments show that the method can smooth the robot trajectory in real time and reduce the pause.

Secondly, the evaluation index of parameter calibration method is proposed for the dynamic parameter calibration problem. The commonly used inertial parameter and friction parameter calibration models are compared, and the advantages and disadvantages of different methods are analyzed. By simulations and experiments, the results of different calculation methods are compared, and the selection strategy of the model and calculation method is illustrated by taking the six-degree-of-freedom industrial robot ER20-C10 as an example.

Thirdly, the external force estimation and force following control method of the multi-joint robot are proposed to deal with the sensorless guiding problem. The inertial force and friction of the links and driving systems are analyzed. The nonlinear dynamic model of the industrial robot is built and its parameters are calibrated by an optimization method. A guiding method based on virtual mass and virtual friction is proposed to make the end-effector follow the direction of external force and suppress jitters. The experimental results show that the calibrated parameters can accurately calculate the external force and the control method achieves good guiding performance. 

Fourthly, for the design and control problem of the collaborative manipulator, the structure and control system of the robot are designed, the inertia of links is reduced, and the path following method under constrained joint torques and the trajectory tracking control based on the feedback linearization method are analyzed. The safety of human-robot collaboration is improved and the tracking error is reduced. Finally, the proposed trajectory planning and dynamic control methods are integrated into the robot system. Experiments verify the effectiveness of the proposed methods and designs.

Finally, the conclusions are given and the future work is addressed.