车辆行驶过程中，多诱因引发的驾驶人异常情绪对车辆安全影响极大，因而将驾驶人异常情绪引入到智能车辆监控范围中来，具有重要的现实意义。然而，目前对驾驶人异常情绪与人-车-环境耦联系统间的作用机理和映射关系还不十分清楚，对驾驶人异常情绪的判断标准、表征方法和在线识别等方面的研究仍有待深入。为此，本项目以可能引发车辆危险的驾驶人异常情绪为研究对象，以实现车辆安全态势早期估计为出发点，以驾驶人异常情绪的内涵界定、特征筛选、反演辨识为主要研究内容。采用理论建模与实验验证相结合的方法，探索非接触式、个性化、可在线主动学习的驾驶人异常情绪反演辨识方法。旨在明确可能导致车辆危险的驾驶人异常情绪内涵与界定法则，阐明人-车-环境耦联系统中非接触式多维异构特征与驾驶人异常情绪间的内生联系与映射规律，为构建高可信的车辆安全态势早期评估模型提供理论依据和前沿技术支撑。

When driving a vehicle on the road, abnormal emotions triggered by multiple factors may have a serious impact on vehicle’s safety. Thus, taking account of the influence of the abnormal emotions into vehicle safety assessment has a practical significance. However, currently, the mapping relationship between abnormal emotions and human-vehicle-environment system is not very clear and the criteria, characterization, and online identification of abnormal emotion remains to be elucidated. Therefore, this project will focus on the driver’s abnormal dangerous emotion and study the early assessment method of vehicle's safety. In order to achieve the goal, the project will take advantage of both theoretical modeling and experimental verification method, with the main research content of abnormal emotion definition, feature selection and inversion identification. This research intend to make clear of the definition rule of abnormal emotion and clarify the endogenous contact between multi-features of driving system and driver’s abnormal emotion. Then, the method can provide both theoretical and technical support for developing an early assessment model of vehicle’s safety.