Deep-Learning-Based Deconvolution of Mechanical Stimuli with Ti3C2Tx MXene Electromagnetic Shield Architecture via Dual-Mode Wireless Signal Variation Mechanism

Abstract

Passive component-based soft resonators have been spotlighted in the field of wearable and implantable devices due to their remote operation capability and tunable properties. As the output signal of the resonator-based wireless communication device is given in the form of a vector (i.e., a spectrum of reflection coefficient), multiple information can, in principle, be stored and interpreted. Herein, we introduce a device that can deconvolute mechanical stimuli from a single wireless signal using dual-mode operation, specifically enabled by the use of Ti3C2Tx MXene. MXene's strong electromagnetic shielding effect enables the resonator to simultaneously measure pressure and strain without overlapping its output signal, unlike other conductive counterparts that are deficient in shielding ability. Furthermore, convolutional neural-network-based deep learning was implemented to predict the pressure and strain values from unforeseen output wireless signals. Our MXene-integrated wireless device can also be utilized as an on-skin mechanical stimuli sensor for rehabilitation monitoring after orthopedic surgery. The dual-mode signal variation mechanism enabled by integration of MXene allows wireless communication systems to efficiently handle various information simultaneously, through which multistimuli sensing capability can be imparted into passive component-based wearable and implantable electrical devices.