面向车联网的道路交通事故链动态演变规律及其阻断方法研究

项目名称： 面向车联网的道路交通事故链动态演变规律及其阻断方法研究

项目编号： No.61573171

项目类型： 面上项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 陈龙

作者单位： 江苏大学

项目金额： 16万元

中文摘要： 以道路交通连环事故为研究对象，探究交通事故与交通事件间的关联关系，针对道路交通事故链(CRTI)在时间序列上满足马尔可夫链的“无后效性”特征，基于马尔可夫理论，揭示CRTI生成与演化机理和规律，建立CRTI演化模型，将复杂多样的连环事故表征成有序的CRTI形态，掌握不同的控制参数对事故危害结果的影响规律，提出机器学习和交通元事件搜索匹配方法，构建“交通情景-应对”应急决策新范式和CRTI监控算法，以跟踪CRTI的演化趋势，实现CRTI的有效监控。针对连环交通事故中目标车辆具有的邻接影响、条带状传播特性，设计不同危险度和危险源的系列事故链阻断策略，建立基于车联网上下文感知的阻断效率动态保障机制与评价方法,实现事故“链”和进程的有效中断。最后，建立CRTI阻断原型系统，以验证有效性。该项目旨在有效控制CRTI形成、降低连环交通事故损失和伤害，为道路交通安全主动防控提供新的理论依据和技术支持。

中文关键词： 交通安全；交通事故链；阻断；演化；车联网

英文摘要： The project takes serial traffic accidents as the research target, investigating the correlation of traffic accidents and traffic events. Focusing on the non-after-effect property in time series of chain of road traffic incidents (CRTI), it reveals the generation and evolution mechanism and regularity of CRTI and establishes the evolution model of CRTI based on Markov chain so as to represent various serial traffic accidents as ordered CRTI forms. It analyzes the influence regularity of different control parameters on accidental harmful results and proposes the machine learning and traffic meta-events searching and matching algorithms. It constructs new “traffic situation – reaction” emergency decision-making scheme and CRTI surveillance algorithm to track the evolution trend of CRTI and monitor CRTI effectively. Focusing on the neighboring influence and banding propagation character of target vehicle involved in serial traffic accidents, it designs different CRTI blocking strategies for various dangerous degrees and sources. It also establishes the dynamic guarantee mechanism of the blocking efficiency and evaluation method based on the context perception in Internet of Vehicles conditions, which can effectively block the accident chain and process. Finally, the project constructs a CRTI prototype system to ver

英文关键词： Traffic safety；chain of road traffic incidents；blocking approaches；dynamic evolution regularity；Internet of Vehicles