A Strain-Sensitive Flexible MoTe2-Based Memristor for Gesture Recognition

Conventional flexible electronics suffer from complex fabrication processes and great energy consumption due to the separated tactile sensors, storage devices, and processing units. In-sensor computing based on multifunctional devices allows efficient computation by directly sensing and processing signals and is considered an effective way to achieve text large-scale integration and low-power devices in the future. Herein, a strain-sensitive multifunctional memristor is prepared based on MoTe2. The device shows good retention (>105 s) and switching stability (>500 cycles). The neuromorphic computing capability of the device was verified using a gesture recognition task with accuracy reaches above 70% even at 50% background noise. The present results demonstrate the attractive multifunction integrated behaviors of MoTe2 devices, opening a new avenue for designing artificially intelligent electronic skins.