【作者】 魏媛媛；

【导师】 朱绍文；

【作者基本信息】 华中师范大学 ， 计算机应用技术， 2001， 硕士

【摘要】 智能化是自动化的最高表现，机器人的智能化研究已成为科学技术发展的一个主要方向。对智能机器人来讲，主要有这样2个核心内容：一是在行动中系统对环境的感知；二是感知后机器人的实时行动，感觉和行动的高度统一是智能机器人研究的必然要求。因此，智能机器人力觉和力控制研究是智能机器人基于感觉和行动综合研究的重要组成部分，属于智能机器人研究的基础。 力控制研究是力觉研究之核心，力觉和力控制贯穿于智能机器人作业的全过程。力控制的典型特点是力/位置强耦合，是一种极其复杂的混合控制，与智能机器人研究紧密联系着，一直是研究的主要方向。 近二十年来，机器人装配作业逐渐成为科学家们研究的热点。它涉及自动化技术、机构学、传感器技术、材料科学、信息处理等多种学科，机器人力控制是其中最基础和核心的内容。 机器人装配作业是自动柔性制造系统的关键环节，它要求高精度的位置信息和装配工件的精确几何特征。机器人装配作业的研究对于扩大机器人的应用范围，有极其重要的意义。在机器人装配操作中，由于装配任务的不确定性（包括产品公差、控制不确定性和传感误差等），单纯的位置控制机器人不能胜任带有接触约束的装配任务，即使是较小的位置误差都可能引起工件与环境之间较大的接触力而损坏机器人或装配体。如何使工业机器人实现快速、精密的装配作业是目前尚未完全解决的问题。 在实际装配过程中，由于装配件尺寸、定位、传感信息以及机器人运动所造成的不确定性的存在，使得装配过程极易发生故障。因此，许多学者对装配过程进行了大量的卓有成效的研究。由于机器人装配实质是一系列接触状态的变迁过程，装配控制策略应随装配状态的不同而进行调整，才能进行快速有效的控制。 本文主要介绍了国内外的最新的机器人柔顺控制方法，并对其进行了较为详细分析、研究。指出了机器人柔顺控制方法中存在的关键问题以及解决这些问题的途径，并指出了今后研究的方向。 在此基础上，本文提出了一种基于模糊控制理论的机器人柔顺控制方法，并利用此方法在Motoman UP6机器人上进行了机器人装配实验，实验结果表明了本方法的正确性和有效性。同时提出了一种基于前馈迭代的机器人柔顺控制算法，对此算法进行了仿真研究，得出0；自霎霎ZI 0十9什个了t：自自自／／J”人 上 丁B－－－x卜乙 八＼‘。”、－八1u乃RS 了HESIS 了仿真结果，对机器人柔顺控制算法的进一步研究有借鉴作用。更多还原

【Abstract】 The uppermost represent of automatization is intelligentize. The study of robot intelligentize has become one main aspect of technology development. As for intelligentize robot, there are 2 cores: the one is the percipience to environment of the system in those actions; the second is the robot real time action after the percipience. The highly unification of these two aspects is necessity requirement. Therefore, the study in force sense and force control is important component on comprehensive research in sensibility and action of robot. It is belong to the foundation of robot study. Force control is the core of force sense research. Force sense and force control interpenetrate the whole process of the robot performance. The typical feature of force control is force/position strongly coupling, which are very complex hybrid control and the main research direction of robot study. In the last twenty years, robot assembly is the key link of the auto- flexibility manufacture. It requires high-precision position information and the precise geometric feature of the assembling work piece. It is very important of the research of robot assembly to enlarge its application rang. In the robot assembly, because of the uncertainty of assembly, such as product common difference, controlling uncertainty and sensor error etc.), simply position control can抰 implement the assembly task with touch constraint. Since the little position error will possibly cause large touch force between work piece and environment, which can be damaged to robot or assembly body. It is not complete solution that how to quickly and precisely assembly. In the practice, because there are some uncertainties of assembly body, location, sensor error and robot motion, the assembly process might occur accident. Many researchers have done many good jobs. Since robot assembly is the historical development process of serial of touch states, the control strategy of assembly must adjust with the different states, which might have quick and effective control. In this thesis, the latest robot compliant control methods are introduced. And the detailed analysis and research of these methods are presented. Ill The key problem of robot compliant control and the solution are pointed out. The research directions are also pointed out On the bases, one method based on fuzzy control robot compliant control is presented. And the assembly experimentation is preformed using this method on Motoman UP6 robots. This experimentation shows .1 the method is correct and effective. At the same time, one algorithm based on feedforward iteration robot compliant is presented. And the simulation result may be a reference to others.更多还原