面向人机共融协作的柔性双臂机器人安全作业机理研究

The fading of demographic dividend has become the prominent restriction factor of development of new manufactory industry. On account of poor security and flexibility, human-robot close collaboration cannot be achieved with traditional industrial robot, which limits the promotion of robot in new manufactory industrial like 3C. Two basic scientific issues, robot flexibility and security operating, become the top priorities and the main difficulties for realizing of human-robot close collaboration. In this project, these two issues are to be investigated. Firstly, basing on mechanism analysis of human-joint variable stiffness, the design method of a new active-switching-style flexible bionic joint basing on angle feedback is to be developed. Then, the flexible control of dynamic-parameters-optimization-characterized personification robot joint can be achieved. Secondly, for realization of safety human-robot collaboration, the mapping model of collision forces and joint forces is to be developed basing on transmission mechanism of manipulator joint force. Then a new mothed is to be investigated for online estimation of collision force and location of collision position with non-perfect sensor information under non-predictive collision. Thirdly, for human-robot-work collaboration, the coordinated control method of dual-arm system is to be studied basing on relative dynamic, where operator and external environment are chosen as constraints. Finally, a dual-arm robot system is to be developed for experimental study. The theory and technology such as robots natural flexibility and intrinsic safety developed in this project would lay the foundation of dual-arm robot promoting.

单纯以“人”或“机”为主的生产方式难以满足3C等新兴制造业“柔性制造”需求，人机紧密协作将成为重要发展趋势，而传统工业机器人因柔性与安全性问题导致人机空间相隔离，因此解决机器人柔性、安全性作业两大基础科学问题成为实现人机紧密协作的首要任务和主要难点。针对以上问题，本项目首先基于人体关节变刚度机理分析，提出基于主动切换式角度反馈的柔性仿生关节设计方法，实现具有拟人关节动力学参数优化特征的关节柔性控制；其次，基于机械手关节力传递机制建立碰撞外力与各关节力映射模型，提出基于非完备传感信息下的非预测性碰撞在线外力估计与碰撞定位方法，实现人机安全作业；再次，面向操作任务，以操作者及外部环境为约束条件，提出基于相对动力学的双臂机器人系统协调控制方法，实现人机物协调作业；最后，构建类人双臂机器人并进行实验研究；课题获得的机器人自然柔顺和本质安全等方面的理论和技术，可为双臂机器人应用奠定理论和技术基础。

本项目围绕双臂机器人的安全作业与柔顺协调作业的基础理论和关键技术，所涉及的主要研究内容和结果如下：.（1）为了解决机器人关节柔性问题，项目借鉴了生物关节天然卓越的柔性运动特征，提出人体非线性动力学模型时变参数估计方法。通过切换式角度反馈来研究关节柔性生成机理。通过仿生特征参数来优化仿生关节柔性特征。从而将生物关节的运动规律应用到电机驱动的仿生关节。.（2）为了实现机器人安全作业问题，建立了基于碰撞力学和动力学来研究机械手外力扰动动力学模型，从而建立预测性碰撞下的外力与关节柔性力映射模型。并基于关节柔性力与力传感器和融合视觉信息的全工作空间来研究非预测性碰撞规则，从而实现机器人非预测性碰撞实时在线监测。.（3）项目建立了基于相对动力学的双臂机器人建模，提出了基于相对动力学模型的双臂协调控制方法，实现了在同一动力学框架下对面向任务的双臂机器人多约束条件下的构型进行优化。.（4）为验证提出的仿生关节柔性生成机理、非预测性碰撞模型和双臂协调控制方法，本项目基于虚拟样机技术优化双臂机器人的结构参数，对现有双臂机器人系统进行优化。并对仿生关节柔性机理、非预测性碰撞和协调控制方法进行实验验证。.以上研究内容和结论为双臂机器人系统的发展提供理论与技术方案之外，还将为解决其他机器人领域研究中的难题提供可能的方法，具有重要的研究意义。