Research on Fatigue Data Collection Methods for Civil Aviation Cockpit Crews

Abstract

The single pilot operation (SPO) mode has become a promising solution to address the critical lack of qualified pilots and rising costs in the civil aviation industry. However, this operational mode introduces challenges in pilot fatigue management, as the single pilot must afford the entire operational workload without the replacement and cross-monitoring traditionally provided by a co-pilot. Therefore, pilots’ fatigue detection and management are becoming important. This paper systematically reviews current methodologies for pilot fatigue data collection, providing a comparative analysis of physiological signal monitoring techniques, including electroencephalogram, electrocardiogram, and electrooculogram, and non-contact image-based behavioral analysis technologies. This paper proposes a multimodal fusion framework that combines data from wearable physiological sensors with computer vision-based behavioral monitoring. Comprehensive system architecture encompasses four dedicated layers: multimodal data acquisition, advanced signal processing, intelligent feature fusion, and dynamic alerting. By leveraging machine learning algorithms, the system enables real-time, accurate fatigue assessment with enhanced robustness. This multimodal strategy effectively mitigates the constraints of single-method systems, providing a reliable and scalable technological foundation for ensuring aviation safety in the evolving SPO environment.