考虑驾驶负荷的认知决策行为建模及驾驶绩效改善方法

Drivers' Errors on cognition, decision-making and operation are considered as important factors that lead to road accidents in high workload and complex environments. However, the mechanism of driver’s cognitive processes and decision-making with high driving workload in complex environment remains unclear, and the prevention of accident is not efficiency. The project aims to identify the cognitive and decision-making mechanism of the driver. Firstly, complex environments environmental characteristics of accident-prone sections are analyzed based on the actual highway accident data statistical analysis; after driving scene and driving tasks design, driving simulator is employed as a platform to carry out experiments to test the drivers' psychology, and the data synchronization technology is used to collect all kinds of data; pattern recognition methods will be used to identify the driver’s cognitive and behavioral characteristic parameters in a variety of driving loads, also driving character change rules will be recognized. To study the process and mechanism of cognition, ACT-R modeling technology will be implemented to build the driver's cognitive model in the high- workload environment, and HMM will be used to model driver's decision-making process in high-load environment. It will identify the driving performance impairment and its main cause based on driving behavior and decision-making output from the model. At last, traffic safety improvement measures from the aspects of driving behavior improvement and road conditions improvement will be promoted. The research work can complement and improve the existing traffic safety and driving behavior theories, and provide support for the analysis and effective prevention of mechanism of accident under the high-workload environment.

驾驶人在高负荷、复杂环境中的认知、决策和操作失调是导致道路交通事故的重要因素；然而，驾驶负荷环境下的驾驶认知过程及决策机理尚不明确，事故分析预防存在瓶颈。课题从驾驶负荷和复杂环境下的驾驶人认知和决策行为机理出发，基于实际道路交通事故数据统计分析，初步识别事故多发路段环境特征；设计不同道路环境和驾驶任务，以驾驶模拟器开展交通心理学实验，采用数据同步技术采集各类数据；通过模式识别方法提取驾驶人在各种驾驶负荷下的行为特征参数，辨识其演变规律；采用ACT-R认知结构体系建立考虑负荷因素的驾驶人认知行为模型，同时采用HMM建立负荷环境下驾驶人的决策过程模型；针对模型输出的驾驶行为和绩效指标，分析其驾驶绩效损伤特性和主要因素；从道路交通环境和驾驶行为干预两个方面提出交通安全改善方法。课题研究成果可补充和完善现有的驾驶行为量化分析相关理论，为驾驶负荷环境下的事故机理分析和有效预防提供支撑。

驾驶人在高负荷、复杂环境中的认知、决策和操作失调是导致道路交通事故的重要因素；然而，驾驶负荷环境下的驾驶认知过程及决策机理尚不明确，事故分析预防存在瓶颈。课题从驾驶负荷和复杂环境下的驾驶人认知和决策行为机理出发，基于实际道路交通事故数据统计分析，初步识别事故多发路段环境特征；设计不同道路环境和驾驶任务，以驾驶模拟器开展交通心理学实验，采用数据同步技术采集各类数据；通过模式识别方法提取驾驶人在各种驾驶负荷下的行为特征参数，辨识其演变规律；采用ACT-R认知结构体系建立考虑负荷因素的驾驶人认知行为模型，同时采用HMM建立负荷环境下驾驶人的决策过程模型；针对模型输出的驾驶行为和绩效指标，分析其驾驶绩效损伤特性和主要因素；从道路交通环境和驾驶行为干预两个方面提出交通安全改善方法。项目执行期间共发表了21篇国内外学术期刊论文和11篇国内外会议论文，申请了8项发明专利，其中1项专利已授权。项目研究成果可补充和完善现有的驾驶行为量化分析相关理论，为驾驶负荷环境下的事故机理分析和有效预防提供支撑。