Templated Laser-Induced-Graphene-Based Tactile Sensors Enable Wearable Health Monitoring and Texture Recognition via Deep Neural Network

Abstract

Flexible tactile sensors show great potential for portable healthcare and environmental monitoring applications. However, challenges persist in scaling up the manufacturing of stable tactile sensors with real-time feedback. This work demonstrates a robust approach to fabricating templated laser-induced graphene (TLIG)-based tactile sensors via laser scribing, elastomer hot-pressing transfer, and 3D printing of the Ag electrode. With different mesh sandpapers as templates, TLIG sensors with adjustable sensing properties were achieved. The tactile sensor obtains excellent sensitivity (52260.2 kPa^–1 at a range of 0–7 kPa), a broad detection range (up to 1000 kPa), a low limit of detection (65 Pa), a rapid response (response/recovery time of 12/46 ms), and excellent working stability (10000 cycles). Benefiting from TLIG’s high performance and waterproofness, TLIG sensors can be used as health monitors and even in underwater scenarios. TLIG sensors can also be integrated into arrays acting as receptors of the soft robotic gripper. Furthermore, a deep neural network based on the convolutional neural network was employed for texture recognition via a soft TLIG tactile sensing array, achieving an overall classification rate of 94.51% on objects with varying surface roughness, thus offering high accuracy in real-time practical scenarios.