Emotion-induced traffic accidents often arise from a temporary depletion in emotional awareness management (EAM). Technology-assisted EAM systems (TEAMS), which combine real-time emotion recognition with timely regulation strategies, offer a promising approach to mitigating emotion-related driving risks. This systematic review synthesizes current research on emotion recognition and emotion regulation in driving contexts. We conducted a structured search across ACM Digital Library, IEEE Xplore, Scopus, and Web of Science, followed by screening based on predefined criteria. A snowballing search was also used to identify studies on emotion regulation. Ultimately, 134 peer-reviewed studies were included. The results revealed an isolated, unbalanced development of emotion recognition and emotion regulation, with systems that integrate both remaining scarce. Existing emotion recognition research focused predominantly on facial and physiological cues, followed by speech and driving features. Most studies emphasized algorithmic performance and relied on datasets outside driving contexts, which limited ecological validity and generalizability. Research on emotion regulation remained nascent, with studies exploring regulatory strategies such as auditory, visual, and combined feedback. Most of these studies were conducted in laboratory settings, and evaluation approaches varied, often relying on questionnaires or physiological measures. This review suggests the need for future efforts to develop unified, adaptive, and human-centered TEAMS. It also recommends creating diverse, accessible multimodal driving datasets and establishing comprehensive evaluation frameworks that cover objective and subjective measures. Human-centered TEAMS may reduce emotion-induced accidents and enhance safety and interaction during transitions to higher levels of driving automation, thus supporting the development of future intelligent transportation systems.
扫码关注我们
求助内容:
应助结果提醒方式:
