• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2026, Vol. 62 ›› Issue (11): 241-248.doi: 10.3901/JME.260162

• 机器人及机构学 • 上一篇    

扫码分享

面向智能机器人的大面积电子皮肤感知系统

张聘1,2, 姜山1,2, 韩成3, 白龙1,2, 布天昭3, 吴豪3   

  1. 1. 航空工业济南特种结构研究所高性能电磁窗航空科技重点实验室 济南 250023;
    2. 电磁功能结构技术创新中心 济南 250023;
    3. 华中科技大学机械科学与工程学院 武汉 430074
  • 收稿日期:2024-11-12 修回日期:2025-09-28 发布日期:2026-07-29
  • 作者简介:张聘,男,1983年出生,研究员,副总工程师。主要研究方向为先进航空制造技术;E-mail:sdzhangpin@sohu.com;姜山(通信作者),男,1993年出生,博士。主要研究方向为先进航空制造技术,电子封装技术。E-mail:jiangs060@avic.com

Large-area Electronic Skin Sensing Systems for Intelligent Robots

ZHANG Pin1,2, JIANG Shan1,2, HAN Cheng3, BAI Long1,2, BU Tianzhao3, WU Hao3   

  1. 1. AVIC Research Institute for Special Structures of Aeronautical Composite, Aviation Key Lab of Science and Technology on High Performance Electromagnetic Windows, Ji'nan 250023;
    2. Innovation Centre for Electromagnetic Functional Structure, Ji'nan 250023;
    3. School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074
  • Received:2024-11-12 Revised:2025-09-28 Published:2026-07-29

摘要: 电子皮肤为智能机器人提供类似于人类的触觉感知功能,是实现人机交互与灵巧操作的关键技术之一。尽管现有电子皮肤已具备一定的传感功能,但其传感区域较小且缺乏信号处理能力,无法满足机器人对大面积智能化触觉感知的需求。针对上述问题,开发了一种大面积电子皮肤感知系统,并探索了其在智能机器人领域的应用。该感知系统由电子皮肤、数据采集与处理装置组成,具有可扩展的传感面积和高空间分辨率,能够实时反映外部压力分布状态。并且受生物整体感知启发,提出了一种自监督传感校正网络,利用时空上下文信息从弱监督中学习,以解决大面积电子皮肤难以标定的问题。此外,将电子皮肤感知系统应用于触摸模态识别和机器人碰撞检测,拓展了其应用场景,表现出了在智能机器人、人机交互等领域的潜在应用价值。

关键词: 电子皮肤感知系统, 自监督传感校正, 触摸模态识别, 碰撞检测

Abstract: Electronic skin is one of the key technologies to achieve human-computer interaction and dexterous operation, providing intelligent robots with human-like tactile sensing functions. Existing electronic skins already have certain sensing functions, but their sensing areas are small and the signal processing capabilities are not satisfying, making them unable to meet the robot’s needs for large-area intelligent tactile perception. In response to the above problems, a large-area electronic skin sensing system has been investigated and preliminarily applied in intelligent robots. The sensing system consisting of electronic skin, data acquisition, and processing devices, has an expandable sensing area and high spatial resolution which can reflect the external pressure distribution status in real time. Inspired by biological holistic perception, this research proposes a self-supervised sensing correction network using spatiotemporal context information to learn from weak supervision to solve the problem of difficult calibration of large-area electronic skins. In addition, this research applies the electronic skin sensing system to touch mode recognition and robot collision detection, expanding its application scenarios and showing its potential application value in fields such as intelligent robots and human-computer interaction.

Key words: electronic skin sensing system, self-supervised sensing correction, touch modality recognition, collision detection

中图分类号: