Epidermal piezoresistive structure with deep learning-assisted data translation

Abstract Continued research on the epidermal electronic sensor aims to develop sophisticated platforms that reproduce key multimodal responses in human skin, with ability sense various external stimuli, such as pressure, shear, torsion, and touch. The development of applications utilizes algorithmic interpretations analyze complex stimulus shape, magnitude, moduli epidermis, requiring multiple equations for attached sensor. In this experiment, we integrate silicon piezoresistors a customized deep learning data process facilitate precise evaluation assessment stimuli without need complexities. With surpass conventional vanilla regression models, classification model is capable predicting magnitude force, hardness object shape an average mean absolute percentage error accuracy <15 96.9%, respectively. technical learning-aided consequent accurate provide important foundations future sensory system.