Deep Learning Assists Triboelectric Sensor for Emotion Classification

Abstract

Emotions play a crucial role in influencing human behavior and decision-making processes. Accurate recognition of emotional states not only is fundamental to understanding human psychology but also serves as a crucial enabler for applications such as mental health monitoring, human–computer interaction, and intelligent systems. Triboelectric nanogenerators (TENG) have gained significant attention as efficient and wearable energy-harvesting devices with exceptional potential in the sensing domain. The main objective of this study is to classify the emotions of TENG electrical signals. To achieve this goal, an eye-integrated triboelectric nanogenerator sensor was designed, which is capable of converting the mechanical energy generated by micromovements of facial expressions into electrical signals. The positive and negative triboelectric layers of TENG are nylon film and polydimethylsiloxane (PDMS) film. The sensor consists of two TENGs connected in series. Meanwhile, the bidirectional long- and short-term memory network incorporating the attention mechanism has been proposed. When combined with the TENG, the emotion categorization system achieves an accuracy of 94%. The proposed system is demonstrated to have high accuracy in recognizing emotional states, providing a practical and reliable solution for emotion recognition. This study showcases triboelectric nanogenerators’ potential in wearable sensing and emotion recognition applications.