High-Performance MXene/TPU Flexible Piezoresistive Sensors for Enhanced Gesture Recognition and Interactive Control Applications

Abstract

Recent advancements in human-machine interaction have demanded flexible sensors with high sensitivity and mechanical robustness for real-time gesture recognition. In this study, a novel MXene/TPU-based flexible piezoresistive sensor is developed through a combination of electrospinning and MXene functionalization techniques. By optimizing the soaking–drying cycles of MXene nanosheets on TPU nanofiber membranes, the sensor achieved tunable sensitivity. The sensor demonstrated an exceptional sensitivity of 78 556 within the strain range of 110–150% and exhibited outstanding durability (>3000 cycles). Furthermore, a smart glove embedded with five sensors is developed, integrated with a low-power multichannel acquisition system and an LSTM–Transformer hybrid model. The integrated system delivers superior gesture recognition accuracy of 99.47% across 32 gestures, outperforming conventional methods such as SVM (96.47%), KNN (93.51%), and RNN (98.77%). In addition, real-time robotic hand control and VR interaction further validated its practicality in wearable electronics. Overall, this study presents a comprehensive framework for the design of flexible sensors with enhanced sensitivity, stretchability, and stability, contributing to the advancement of intelligent human-machine interface technologies.