Li Nan, Zhan Fei, Guo Minghui, Yuan Xiaohong, Chen Xueqing, Li Yuqing, Zhang Guangcheng, Wang Lei, Liu Jing
State Key Laboratory of Cryogenic Science and Technology, Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, Beijing, 100190, China.
School of Future Technology, University of Chinese Academy of Sciences, Beijing, 100049, China.
Adv Mater. 2025 May;37(19):e2419524. doi: 10.1002/adma.202419524. Epub 2025 Mar 26.
The advancement of robotic behavior and intelligence has led to an urgent demand for improving their sensitivity and interactive capabilities, which presents challenges in achieving multidimensional, wide-ranging, and reliable tactile sensing. Here an anisotropic inductive liquid metal sensor (AI-LMS) is introduced inspired by the human fingertip, which inherently possesses the capability to detect spatially multi-axis pressure with a wide sensing range, exceptional linearity, and signal stability. Additionally, it can detect very small pressures and responds swiftly to prescribed forces. Compared to resistive signals, inductive signals offer significant advantages. Further, integrated with a deep neural network model, the AI-LMS can decouple multi-axis pressures acting simultaneously upon it. Notably, the sensing range of Ecoflex and PDMS-based AI-LMS can be expanded by a factor of 4 and 9.5, respectively. For practical illustrations, a high-precision surface scanning reconstruction system is developed capable of capturing intricate details of 3D surface profiles. The utilization of biomimetic AI-LMS as robotic fingertips enables real-time discrimination of diverse delicate grasping behaviors across different fingers. The innovations and unique features in sensing mechanisms and structural design are expected to bring transformative changes and find extensive applications in the field of soft robotics.
机器人行为和智能的进步引发了对提高其灵敏度和交互能力的迫切需求，这在实现多维、广泛且可靠的触觉传感方面带来了挑战。在此，受人类指尖启发引入了一种各向异性电感式液态金属传感器（AI-LMS），它天生具有在宽传感范围内检测空间多轴压力的能力，具有出色的线性度和信号稳定性。此外，它能够检测非常小的压力，并对规定的力迅速做出响应。与电阻信号相比，电感信号具有显著优势。此外，与深度神经网络模型集成后，AI-LMS 能够解耦同时作用于其上的多轴压力。值得注意的是，基于 Ecoflex 和 PDMS 的 AI-LMS 的传感范围分别可以扩大 4 倍和 9.5 倍。为了进行实际演示，开发了一种高精度表面扫描重建系统，能够捕捉三维表面轮廓的复杂细节。将仿生 AI-LMS 用作机器人指尖能够实时区分不同手指的各种精细抓握行为。传感机制和结构设计方面的创新和独特特性有望带来变革性变化，并在软机器人领域找到广泛应用。
