We develop an ultra-sensitive skin-like sensor that discovers previously undetectable signals far from the joint, assisted by a deep neural network that deciphers patterns within the signals. By utilizing a single sensor capable of extracting signals from multiple areas rather than pinpointing every joint and muscle, we avoid impracticality and provide a more efficient method to understanding the human body.We also propose a concrete theoretical model which discloses the relation between sensor performance and the laser-induced nanoscale crack junctions, thereby facilitating manipulation. Laser fabrication enables a focused sensor patterning, with the sensor conformably attached to the epidermis while achieving high sensitivity. A single deep learned sensor decode finger motions in a real-time demonstration with a virtual 3D hand that mirrors the original motions. This work is collaboration with Prof. Seung Hwan Ko’s group from Mechanical Engineering at Seoul National University.
1. KK Kim, I Ha, M Kim, J Choi, P Won, S Jo, SH Ko, A deep-learned skin sensor decoding the epicentral human motions, Nature Communications, 11: 2149, 2020 [LINK]