协作机器人视—力觉反馈控制系统研究

In industrial occasions such as assembling of electronic components and mechanical parts, it is very difficult to finish all tasks by robots automatically. Human-robot collaboration allows robots to better coordinate with workers and to adapt more complex tasks. Human-robot collaboration has become the key trend for the development of the next new generation of robots. However, research on collaborative robots has just started at present, and most of the research results are published in the last five years. Many basic and key issues have not been solved yet. For example, the theoretical model of multi-interactive system integrating human, robots and environment has not been established, the ability of human and environment perception of robots is still lacking, and collaborative robot's security strategy and motion control algorithms need to be solved. ..This project proposes a collaborative robot control system based on visual and force feedback, which will be studied according to the theoretical analysis, algorithms simulation, and experimental verification technology route. It will establish a human-robot-environment interaction dynamics model, and realize the robot’s environment perception by combination of visual, force and other multi-sensor information fusion. Meanwhile, the safety mechanism of human-robot collaboration will be established, and the motion control algorithm of collaborative robot based on visual and force impedance will be studied. Through this project, the intelligent control of collaborative robot will be realized, and the human-robot collaborative task will be achieved safely, accurately and efficiently. It will provide theoretical and technical support for China’s independent innovation in the next new generation of collaborative robots.

在电子器件、机械零部件装配等工业场合，所有的工位完全由机器人自动化实现难度很大，人机协作可以让机器人更好地与工人配合，适应更复杂的工作任务。人机协作已成为新一代机器人发展的重点方向，但关于协作机器人的研究才刚刚起步，大多研究成果集中在最近五年内，很多重要的基础性和关键性问题还未解决，如融合人、机器人、环境的多元交互系统理论模型尚未建立，机器人对人与环境的感知识别还很欠缺，协作机器人的安全策略和运动控制算法也亟待突破。.本项目提出基于视、力觉反馈的协作机器人控制系统，按照理论分析―算法仿真―实验验证的技术路线，建立人-机器人-环境交互动力学模型，通过视觉、力觉等多传感器信息融合实现机器人的工作环境感知，构建人机协同作业的安全机制，研究基于视、力觉阻抗的协作机器人运动控制算法。从而实现协作机器人的智能控制，保证人机协作任务安全、精确、高效完成，为下一代协作机器人自主创新提供理论和技术支撑。

传统生产中绝大部分工业机器人都被安置在静态、确定的“被隔离”环境中，完成重复性的动作，空间上往往由非常坚固的围栏把机器人与人隔离开。近年来，制造业开始出现定制化、个性化、柔性化的发展趋势，这就对传统机器人这种固定的生产模式提出了严峻的挑战。人机协作作为机器人发展的重要方向，要求机器人与人处于同一自然空间，与人自然交互、协调互补，并确保协同任务的安全执行。.本项目面向下一代机器人人机协作的重要发展方向，以人机协作下的机械零部件装配系统为研究对象，研究基于视、力觉反馈的人机协作控制系统，主要开展了以下研究内容：（1）人机协作系统建模，研究机器人与人、环境之间的空间、运动学、力学等多维度变量耦合关系，建立机器人、人与环境的交互系统动力学模型，为人机协作环境下机器人的控制建立动力学模型。（2）机器人对环境的感知识别，基于视觉、力觉、激光等多种传感方式，实现机器人作业环境中的目标识别及跟踪，通过多传感信息融合，得出准确、可靠的人机协作环境描述，为人机协作环境下机器人的控制提供传感基础。（3）协作机器人运动控制策略，建立机器人主动避碰安全策略，针对人机协作装配应用场景，研究基于视觉、力觉的机器人主动柔顺控制算法，实现机器人的实时行为决策与控制。.项目研究成果已初步应用于人机协作装配、乱序工件抓取、搬运等工业自动化场景，为协作机器人控制和人机协作系统构建提供了理论指导。