危险废物
材料科学
摩擦电效应
机器人
卷积神经网络
鉴定(生物学)
机器人学
可穿戴技术
泄漏(经济)
易燃液体
可穿戴计算机
工艺工程
计算机科学
纳米技术
人工智能
跟踪(教育)
液态金属
接口(物质)
汽车工程
实时计算
深度学习
模拟
化学战剂
分辨率(逻辑)
弹道
移动机器人
作者
Guangxiang Gu,Qiheng Liu,Hongwei Gao,Jinyang Zhang,Zhong Lin Wang
摘要
The rapid identification of chemical composition, temperature, and dynamic behavior of leaked liquids by robots during accidental spills of hazardous chemicals would significantly reduce the potential risks to both human health and the environment. Here, we designed a flexible, palm-shaped liquid-sensing e-skin (PSLSES) featuring 128 metal electrodes fabricated via flexible printed circuit (FPC) technology and coated with a fluorinated ethylene propylene (FEP) film. By capturing the local triboelectrification signals along the droplet's trajectory, PSLSES enables multimodal dynamic liquid sensing, simultaneously achieving liquid composition identification, temperature sensing, as well as droplet motion tracking through visualized trajectory patterns. Integrated with a one-dimensional convolutional neural network (1D CNN), PSLSES achieves an ultrahigh identification accuracy of 99.5% across 21 types of liquids, with a high monitoring resolution down to the ppb level and 98% accuracy in identifying liquid temperature. This includes deionized water, acids, bases, salts, and organic solutions, demonstrating broad liquid identification versatility. Compared with previously reported E-skin systems, it achieved faster identification (0.3 s), a lower detection limit (0.1 ppb), broader liquid recognition, and higher accuracy. The integration of PSLSES into wearable robotics systems opens new ways for robots to assist humans in analyzing and handling chemical leakage in hazardous environments.
科研通智能强力驱动
Strongly Powered by AbleSci AI