机器人手最初为模仿人类的上肢系统的结构与部分功能而研发的用于替代人类的抓持、操作以及手势表达功能的一类自动化装置。随着计算机技术、自动化技术、机械生产水平等不断发展，机器人手功能越来越强大，逐渐广泛应用于各个行业。应用水平也从单一的场景、动作逐渐向多模式、情景化、全智能方向发展。这就要求了现如今的机器人手的产品，不仅仅能够实现单一、稳定的动作循环，还应集成可供选择的多种适用于不同条件的操作方式。本文以具有复合抓取模式的欠驱动机器人手为研究对象，旨在实现一个机器人手的多种抓取模式的融合，提高其功能多样性、适应性和稳定性。 本文针对目前欠驱动机器人手的三种基本抓取模式：平夹抓取、耦合抓取和自适应抓取，以及其功能实现方式：齿轮传动式、带轮传动式、连杆传动式、滑块传动式、腱绳驱动式、流体驱动式等，探究其相互融合的方法。针对不同的抓取模式以及不同的实现方式，实现功能复合的方法及难易程度是不同的，且随着融合的抓取模式数量增多，其困难程度加大，甚至难以找到解决方法。本文分析总结各类欠驱动机器人手功能及结构的异同，通过机构原理分析、运动仿真、等效变换，对其进行优化和改进，为多抓取模式机器人手的设计工作奠定了理论基础。 基于以上模式融合方法，探究出将三种基本的抓取模式（平夹模式、耦合模式和自适应模式）融合到一个机器人手中的两种方法：多关节直接串联式融合和两关节可切换并联式融合，并以此为依据设计出两种多模式欠驱动手：三关节平夹耦合自适应机器人手（COPASA手）和电磁切换式多抓取模式机器人手（ESMGM手）。通过机构原理分析、运动空间分析、静力学分析、运动学分析、动力学分析等，综合评价抓取特性，并建立抓取运动评价标准与各相关参数之间的完整映射关系，形成系统化的分析与设计的方法论。 研制出电磁切换式多抓取模式机器人手（ESMGM手）样机，开展抓取模式验证实验和通用抓取实验，通过压力传感器获取抓取力反馈信息，验证ESNGN手可以实现模式切换和模式复合的基础上，分析其对不同目标物体的适应性、抓取运动特性及抓取力的大小和稳定性。实验结果表明ESMGM 手可以实现瞬间切换功能，并能对不同类型的物体完成稳定抓取，抓取空间较广，抓取力分布合理，在日常操作及现代化生产中有广阔的应用前景。
Robot hand was originally developed to imitate the structure and some functions of human upper limb system, which was used to replace human grasping, operation and gesture expression functions. With the continuous development of computer technology, automation technology and mechanical production level, robot hand has become more and more powerful, and is gradually widely used in various industries. The application level also gradually develops from single scene and action to multi-mode, situational and fully intelligent. This requires the current robot hand products, not only to achieve a single and stable action cycle, but also to integrate a variety of optional operation modes suitable for different conditions. In this paper, the underactuated robot hand with compound grasping mode is taken as the research object, aiming to achieve the integration of multiple grasping modes of a robot hand and improve its functional diversity, adaptability and stability. In this paper, three basic grasping modes of underactuated robot hand: parallel grasping, coupled grasping and self-adaptive grasping, as well as its function realization methods: gear driving, belt wheel driving, connecting rod driving, slider driving, tendon rope driving, fluid driving, etc., are discussed to explore their mutual fusion methods. According to different grasping modes and different implementation methods, the ways and the difficulty degree of realizing functional composition are different, and with the increase of the number of grasping modes of robot hand fusion, the difficulty is increased, even it is difficult to find a solution. Based on the analysis of the grasping function and working performance of the traditional underactuated robot hand, this paper analyzes and summarizes the similarities and differences of the functions and structures of various underactuated robot hands, optimizes and improves them through mechanism principle analysis, motion simulation and equivalent transformation, and lays a theoretical foundation for the design of multi grasping modes robot hand. Based on the above mode fusion methods, two effective methods are explored to integrate three basic grasping modes (parallel grasping mode, coupled grasping mode and self-adaptive grasping mode) into one robot hand: multi joint direct series fusion and two joint switchable parallel fusion. Based on this, two kinds of multi-mode underactuated robot hands are designed: three-joint coupled, parallel and self-adaptive robot hand (COPASA hand) and electromagnetic switching multiple grasping modes robot hand (ESMGM hand). Through mechanism principle analysis, motion space analysis, statics analysis, kinematics analysis, dynamics analysis, etc., the grasping characteristics are comprehensively evaluated, and the complete mapping relationship between the grasping motion evaluation standard and relevant parameters is established, forming a systematic analysis and design methodology. A prototype of electromagnetic switching multiple grasping modes robot hand (ESMGM hand) was developed, and the grasping mode verification experiments and general grasping experiments were carried out. The feedback information of grasping forces was obtained by pressure sensor, on the basis of verifying that ESMGM hand can realize mode switching and mode integration, the adaptability of ESMGM hand to different objects, the characteristics of grasping motion, the size and stability of grasping force are analyzed. The experimental results show that ESMGM hand can achieve instant switching, and can stably grasp different types of objects. It has a wide range of grasping space and reasonable distribution of grasping forces. Therefore, ESMGM hand has a broad application prospect in daily operation and modern production.
