触觉不仅是人进行感知与交互的重要生理功能，也是机器人实现精确运动、灵巧抓取和人机交互的重要基础。对人手的触觉感知和灵巧抓取行为的深入研究，将极大启发和促进机器人触觉传感与物体操纵技术的发展。目前，人们对触觉的生理学基础已经有了比较系统的认识，但是对触觉感知与皮肤界面力学的关系的认识却不够深入。从力学角度看，皮肤与物体表面接触与摩擦是人体触觉形成的基础。界面法向和切向接触应力的时间-空间分布特征是触觉感知的本质来源，对接触界面摩擦状态的感知是人手实现灵巧抓取的关键基础。因此，为了构建生物触觉研究与机器触觉研究的桥梁，本工作深入研究界面摩擦行为机理，发展界面接触应力高分辨测量手段，进而揭示人手的摩擦触觉感知机理和灵巧抓取的反馈控制规律，并指导实现了基于触觉传感的机械手灵巧抓取的初步应用。首先，针对人手触觉感知和抓取过程不可避免的水润滑环境，构建了边界润滑状态下润湿性影响粘附和摩擦的定量模型。通过材料表面化学改性的实验设计，基于微观界面力学的理论与实验研究，揭示了表面润湿性影响水润滑行为的本质是范德华吸引力和水合排斥力等多种表面力竞争导致的粘着摩擦现象。该工作不仅加深了对水基边界润滑机理的认识，也为水介质中物体抓取提供了理论指导。其次，针对现有接触应力测量手段在应力表征维度和时空分辨率等方面的不足，提出了一种基于双目立体视觉和弹性力学模型的界面动态三维接触应力高分辨测量方法。搭建的测量装置原型的时间和空间分辨率分别达到了10 ms和10 μm。该方法的准确性通过经典接触力学实验得到验证，方法的优异性能在仿生干粘附表面的粘着应力测量、滚动摩擦粘着阻力和弹性阻力的可视化、蜗牛爬行多尺度吸盘机制的揭示等研究中发挥重要作用。该方法是开展摩擦触觉实验研究和触觉传感设计的技术基础。最后，开展了人手的摩擦触觉感知行为机理和抓取反馈控制策略研究，构建了反映皮肤摩擦的载荷、方向和表面几何参数依赖性的量化模型，系统研究了人手灵巧抓取过程的增量式加载策略和基于界面微滑感知的反馈调控策略。基于上述控制策略和设计的视触觉传感装置，构建了机器人灵巧抓取的触觉反馈控制范式，成功实现了不依赖先验信息的、动态载荷下机械手对未知物体的可靠抓取。

Tactile sensation is not only an important physiological function of human perception and interaction, but also an important basis for robots to realize precise movement, dexterous grasping, and human-machine interaction. The in-depth understanding of the tactile generation mechanism and dexterous grasping of the human hand will greatly inspire and promote the development of robotic tactile sensing and object manipulation technologies. At present, people have a systematic understanding of the physiological basis of tactile sensation, but only have a limited understanding of the relationship between tactile perception and the mechanical behavior at the skin-object interface. From a point of view forces, the contact and friction between the finger skin and surface of an object is the basis for the generation of tactile sensation. The spatio-temporal distribution of the normal and tangential contact stress at the interface is the essential source of tactile perception. The perception of the interfacial friction state is the key basis for a human hand to acieve dexterous grasping. Therefore, in order to build a bridge between the field of biotactile research and machine haptics, this work starts from the revealing the mechanism of interfacial friction, develops a high-resolution measurement method of interface contact stress, and then reveals the mechanism of frictional tactile perception and the feedback control law of dexterous grasping of a human hand, so as to further guide the preliminary application of robotic dexterous grasping based on tactile sensing. Firstly, considering the unavoidable aqueous lubrication conditions in the process of tactile perception and object grasping, a quantitative model of the effect of wettability on adhesion and friction under boundary lubrication was constructed. Based on the theoretical and experimental studies of microscale interfacial forces of chemical modified material surfaces, wettability affected water lubrication behaviort was revealed as an adhesive friction phenomenon caused by the competition of various surface forces including van der Waals attraction and hydration repulsion. The results provide in-depth understanding of mechanism of water-based boundary lubrication and theoretical guidances for the realization of object grasping in water media.Secondly, in order to overcome the shortcomings of the existing interface contact stress measurement methods in terms of stress dimension and temporal-spatial resolution, a dynamic high-resolution measurement method of interface three-dimensional contact stresses based on binocular stereo vision and elastic mechanics model was proposed. The temporal and spatial resolution of the constructed prototype can reach 10 ms and 10 μm, respectively. The accuracy of the method was verified by classical contact mechanics experiments, and the excellent performance of the method played an important role in the researches such as measurement of adhesion stress on dry adhesive surfaces, visualization of the adhesive resistance and elastic resistance of rolling friction, and revealing the multi-scale suction mechanism of snail crawling. This method provides technical support for the tactile friction measurement and tactile sensors design.Finally, the researches on the frictional tactile perception mechanism of the human hand and feedback control strategy of dexterous grasping were carried out. A quantitative model reflecting the load-, direction- and surface geometry-dependence of skin friction behaviors was constructed. Incremental loading strategy and close-loop control based on perceived interfacial micro-slip were systematically studied. Based on the above control strategy and designed image-based sensing devices, a tactile-feedback control paradigm for robotic dexterous grasping was constructed. It was finally realized the robotic reliable grasping of unknown objects under dynamic loadings without prior knowledge.