触觉是人形机器人与具身智能的"第二感官"，对机器人精准交互、灵活作业意义重大。 从2017年起，他山科技便专注于在触觉感知技术领域的探索和研发，如今已成为机器人灵巧手以及整个 具身智能产业链的重要供应商。 近日，他山科技联合创始人兼CEO马扬在接受中国证券报记者采访时表示，随着触觉感知技术的深化应 用，机器人正从"看得见"向"摸得着、懂力度、会适应"进化，这为开拓更多应用场景带来新的可能。比 如，其团队已经训练了机器人预处理小龙虾的雏形方案。"等到明年，送上餐桌的小龙虾可能就是机器 人预处理的。"马扬说。 从底层芯片做起 马扬认为，机器人若想真正成为人类的得力助手，指尖的触觉闭环能力是不可或缺的一块"拼图"。 2017年，他山科技刚成立之时，市面上还没有一家企业能做出可以放进手端边缘端的触觉感知芯片。他 山科技从这颗芯片做起，一做就是四年多。2022年，全球首款人工智能触觉感知专用芯片终于问世。 马扬表示，与视觉器官不同，人类的触觉器官就是皮肤，可以说是最不像器官的器官。"面积大，哪都 能产生数据，通路太多，怎么在底层布置芯片，怎么把多个芯片协同起来，都是我们要思考的问题"。 机器人的视觉只需要看，而触觉涉及 ...

Core Viewpoint - The article discusses the development of a tactile sensing-based shared control strategy for prosthetic hands, which enhances the precision of control for amputees while reducing fatigue, thereby improving the acceptance of myoelectric prosthetic hands [1][2]. Group 1: Tactile Sensing Technology - Tactile sensing technology enables prosthetic hands to detect contact, prevent slipping, and adjust grip strength, assisting users in daily tasks such as pouring water, drinking, and handling objects [1]. - The proposed shared control strategy incorporates multi-stage grasping and force level switching functions to help amputees perform multi-object sorting tasks more efficiently [2]. Group 2: Research Framework - The overall framework of the tactile sensing-based shared control strategy includes three main components: gesture recognition to decode user intent, activation of pre-shaping and stable grasping functions based on contact state detection, and determination of force level switching or multi-stage grasping based on gesture consistency [3]. - The autonomous control module is responsible for determining the activated functions and generating corresponding motor commands based on user intent and fingertip tactile sensing [3]. Group 3: User Study - A user study involved 8 able-bodied participants and 3 upper-limb amputees, focusing on three common daily activities: multi-object sorting, filling a cup with water, and squeezing toothpaste [4]. - Results indicated that the multi-stage grasping function significantly improved operational efficiency (shorter completion time), while the force level switching function enhanced operational quality (higher success rate), effectively reducing muscle usage and receiving positive feedback from users [4]. Group 4: Discussion and Future Directions - The study emphasizes the need for a comprehensive evaluation of new prosthetic hand functions, including task performance, control burden, and usability measurements [5]. - User-centered assessments are established as crucial for evaluating prosthetic systems, highlighting the importance of understanding user willingness to activate and accept new functions in the design of prosthetic hands [5].