From theory to practice: a thorough examination and future-oriented analysis of driver emotion recognition

Driver emotion is a critical factor in traffic accidents, and yet, a cohesive and prospective synthesis of driver emotion recognition research remains elusive. Through a structured analysis of 69 highly relevant studies, we evaluated experimental environments, emotional representation models, datasets, feature extraction methods, and classification approaches across facial, speech, physiological, and multimodal signals to provide a comprehensive review of driver emotion recognition research. Our synthesis reveals the complexities of emotion detection and recognition, presenting a detailed comparative analysis of the current research landscape. The analysis indicates key areas for future research, including the development of more extensive and diverse emotion datasets specific to driving contexts, the enhancement of multimodal signal fusion techniques for improved recognition precision, and the design of non-invasive measurement tools to ensure data reliability and driver safety. Additionally, we emphasize the importance of creating more realistic simulation environments to align research more closely with real-world scenarios. This review serves to offer a clear overview of the current state of driver emotion recognition research and to guide future investigative efforts.