Deep-Learning-Based Prediction of Long-Term Piezoresistive Sensing Performance of MXene/Aramid Nanofiber Sensors

Flexible compressible sensors are widely used in the human health monitoring field for their numerous advantages. However, the dynamic loads and possible injuries associated with long-term living and exercise pose a challenge to the long-term piezoresistive performance stability of these sensors. In this study, the application of deep learning for predicting the long-term performance of these sensors is explored, aiming to enhance the assessment of sensor stability and ensure accurate and reliable long-term monitoring. Samples with different Ti₃C₂T_x MXene/aramid nanofiber mass ratios (1:1, 1:2, 1:3) are prepared and piezoresistive characterization is conducted under long-term loading cycles to obtain training data. Three distinct deep-learning prediction models, convolutional neural network (CNN), long short-term memory, and recurrent neural network (RNN), are utilized to assess their influence on prediction accuracy. To assess the effectiveness of the proposed method, its prediction of long-term piezoresistive sensing performance with experimental data not used for training purposes is compared. The CNN model demonstrates optimal results with a mean absolute error of 0.0251 for the 1:3 mass ratio sample. Based on the experimental results, the model is expected to be integrated into human health monitoring systems, thus improving the assessment of sensor stability throughout its lifetime.