平面交叉口辅助中驾驶认知行为形成机理及其支持方法

In order to solve when, why and how the driver assistance systems will play an important role in urban asafety, this project will be primarily focused on understanding driving cognitive shaping behaviour at signalized and/or unsignalized urban intersections. For this purpose, our research will conduct the following research activities in support of analyzing urban traffic flow on a microscopic level and improving traffic safety. (1) To recognize and identify driving states in a potentially hazardous urban segments composed of simplr and multi-lane segments in cross or T intersections;(2) To analyze driving behavioural and cognitive mechanisms at the rule- - based levels; (3) To specify the characteristics of driver distraction and the mechanisms of visual function; (4) To model the rule-based driving cognitive process combinating various datasets, estimate the parameters, and develop hierarchical cognitive modular architecture by sequencing and mixing pure basic cognitive models; (5) To identify the influence of intersection characteristics on driver attention allocation and driving cognitive behaviour;(6) To evaluate the effection of complex driving states and vice versa on the driving cognitive behavioural shaping process as a traffic benkmaking; (7)To explore driving cognitive error-compensating and safety control strategies for intersection assistance. Our research will contribute to the development of advanced driver assistance systems, and make it possible to improve road capacity and enhance traffic safety.

为了揭示平面交叉口"人失误的不可避免性"的机理及其特征，解决交叉口"什么时候需要驾驶辅助、辅助什么和如何辅助"的问题，本项目通过评测平面交叉口中非机动车、机动车、行人所体现出的复杂交通态势的动态变化，研究规则基驾驶认知行为形成的内在机制及其外部表现方式；通过视觉线索的提取来识别影响驾驶人视野变化的多种交互因素，重点构建能够处理多输入交通信息的规则基驾驶认知行为形成的算法、参数标定及结构模型库；依据交通基准案例来评测驾驶状态意识的层次划分与优化组合对规则基驾驶认知行为形成的影响，确定出基于驾驶认知失误的支持增强内容与控制策略，为实现交通智能化和交叉口辅助提供技术基础与理论依据，以期达到提高道路通行能力和交通安全水平的目标。

本项目主要通过评测平面交叉口中非机动车、机动车、行人所体现出的复杂交通态势的动态变化，研究规则基驾驶认知行为形成的内在机制及其外部表现方式，确定出基于驾驶认知失误的支持增强内容与控制策略。主要研究成果包括：平面交叉口中驾驶认知行为形成的测试与仿真实验，多车道和单车道交叉口视觉基驾驶任务的功能分解，平面交叉口中驾驶人注意力变化的先验知识模型，平面交叉口中驾驶行为形成的认知计算与模型验证。本项目所获得的相关研究成果已分别在国际学术刊物及国际会议上上发表论文20 篇，其中被SCI收录达7 篇、EI收录达11篇，有1篇论文为ESI高被引论文，自2015年间连续3年入选中国高被引学者，主办了绿色智能交通系统与安全国际学术会议4届并做大会报告2次，以第一完成人获得教育部自然科学二等奖和行业科技进步二等奖2项，进行研究生国际联合培养研究生5名。