基于视觉特性的危险品车辆驾驶人驾驶安全性辨识

公路危险品运输已成为交通安全领域亟待解决的关键问题，驾驶人驾驶安全性提升是解决该关键问题的根本途径之一。疲劳驾驶、分心驾驶等不安全驾驶行为会导致驾驶人注意力分散、反应速度变慢等变化，这些变化均能通过驾驶人视觉特征参数进行反映。为此，本项目以驾驶人视觉特性为桥梁，以准确辨识危险品驾驶人驾驶安全性为目标，融合驾驶人个性、行驶道路条件、疲劳驾驶及分心驾驶等信息，采用方差分析、数据拟合、回归分析等方法深度解析视觉特征参数在非常态驾驶过程中的变化规律，并基于多小波理论对参数进行变换，剥离与驾驶安全不相关的其他影响因素，从而提取危险品驾驶人驾驶安全性辨识指标，构建辨识指标体系；在此基础上，基于三角模糊数的模糊分析方法，利用主观与客观相结合的方式，确定辨识指标权重，构建危险品驾驶人驾驶安全状态辨识模型。研究成果将大大减少危险品道路交通事故、丰富交通安全理论体系，并为驾驶安全状态监测提供技术支持。

Road dangerous goods transportation has become a key issue in the field of traffic safety, and driver’s safe driving improvement will be one of the fundamental ways to solve this critical problem. Unsafe driving behaviors, such as fatigue driving and distracting driving, will cause driver's distraction and slow response. These changes can be reflected by the driver's visual characteristic parameters. To this end, this project took the driver's visual characteristics as a bridge, and integrated information such as driver's personality, driving road conditions, fatigue driving and distracting driving to accurately identify the driving safety of dangerous goods transportation. Many methods, including variance analysis, data.preparation and regression analysis, were used to deeply analyze the variation law of visual characteristic parameters in the abnormal driving process, and then transformed the parameters based on muti-wavelet theory to strip other influencing factors that were not related to driving safety, thus extracting driving safety of dangerous goods transportation and constructing identification index system. On this basis, the fuzzy analysis method based on triangular fuzzy number was used to determine the weight of the identification index by combining both subjective and objective methods, and thus to construct the driving safety state identification model of dangerous goods transportation. The research results will greatly reduce road traffic accidents of dangerous goods transportation, enrich the theoretical system of traffic safety, and provide technical support for driving safety monitoring.