A triboelectric sensor array based on facial muscle-action unit-emotion mapping toward astronaut psychological monitoring in long-term missions

In deep space exploration and long-term manned missions, astronauts face significant challenges in operational ergonomics and mental health under sustained stress in isolated, confined, and extreme (ICE) environments. Understanding the “physiological-psychological-task-environment” interaction is crucial for mitigating human-error risks; however, current on-orbit psychological data generally rely on post-mission psychometric scales and lack real-time capability and objectivity. Therefore, developing a long-term, non-invasive emotion monitoring system is essential. Existing visual recognition methods are constrained by lighting and privacy, whereas conventional bioelectrical measurements require cumbersome wet electrodes. In addition, the capability of current flexible sensors to detect subtle facial muscle movements still requires improvement. This study presents a self-powered bionic skin sensor (FAEM) based on a triboelectric nanogenerator (TENG) for continuous, non-invasive facial expression recognition. FAEM employs a multilayer design incorporating a microstructured PDMS film and enables sensitive detection of subtle skin strain by improving interfacial contact regulation and local electrostatic induction. Following the Facial Action Coding System (FACS), the sensing array is arranged over five key facial regions to acquire multichannel signals corresponding to seven typical emotions and a neutral expression. A deep neural network integrating convolutional layers, residual blocks, and Bi-LSTM was constructed, achieving a recognition accuracy of 99.13% and excellent stability under the current evaluation conditions. This study provides a feasible approach for psychological monitoring in extreme environments and demonstrates broad application prospects for future manned space missions.