【作者】 李军；

【导师】 王健；

【作者基本信息】 哈尔滨工业大学 ， 交通运输工程， 2014， 硕士

【摘要】 黑龙江省地处中国最北部，拥有漫长的冰雪期，2008-2011年黑龙江省的高速公里里程实现了跨越式增长，与此同时，黑龙江省的汽车保有量持续增长，近些年高速公路发生的事故也逐年增多，这给黑龙江省的社会、经济都方面带来了诸多的负面影响。冰雪条件下高速公路道路条件急剧下降，加之驾驶员的粗心大意，极易引发交通事故，如何提升冰雪条件下高速公路的安全性能，具有十分重要的意义。本文分析了冰雪条件对于高速公路路面状况、车辆行车的环境能见度、车辆类型以及车辆特性对高速公路交通安全的影响。选取了车辆速度、车辆加速度、车头时距、车辆距左侧车道线距离等4个参数作为特征因子，构建了基于隐马尔科夫模型的冰雪条件下高速公路危险驾驶行为辨识模型，鉴别出超速驾驶，车距不足、违章超车、压线行驶等四种危险驾驶行为。最后，针对这四种危险驾驶行为制订了相应的保障措施，并建立由视频监控子系统，危险驾驶行为辨识子系统，信息发布子系统等3个子系统构成危险驾驶行为实时监控系统。本文首次将常用于语言识别领域的隐马尔科夫模型应用到车辆的危险驾驶行为辨识领域，将该辨识模型嵌入到目前的高速公路监控系统，建立危险驾驶行为实时监控系统。研究表明，基于隐马尔科夫模型的危险驾驶行为辨识模型能够准确地辨别出冰雪条件下高速公路的危险驾驶行为，提出的改善措施也为提升冰雪条件下高速公路安全做出了一定的贡献。更多还原

【Abstract】 Heilongjiang province is situated in the north of China, with long ice and snow time,From2008to2011, highway mileage of Heilongjiang province has seen massive growth.at the same time, the automobile of heilongjiang province has a big increase, the traffic accident rate is rising year by year, has bring deep influences to economy, politics and every aspect of social life of heilongjiang province. Under the condition of ice and snow road, Easily lead to accidents, how to improve the safety performance of highway snow and ice conditions, has the very vital significance.This paper analyzes the snow conditions on the highway road conditions, visibility environment, vehicle type, and vehicle characteristics to traffic on highway traffic safety impact. Select the vehicle speed, vehicle acceleration, distance, vehicle parameters from the four left lane as the front line distance characteristic factor, constructed highway dangerous driving behavior identification model snow conditions based on hidden Markov model, identify speeding driving the car in less, illegal overtaking lane pressure other four dangerous driving behavior. Finally, these four dangerous driving behaviors appropriate safeguards developed and established by the video surveillance subsystem, dangerous driving behavior identification subsystem, information dissemination subsystem three subsystems dangerous driving behavior monitoring system.This paper was first used in the field of speech recognition Hidden Markov model to dangerous driving behavior of the vehicle identification field, the identification model is embedded into the current highway monitoring system to establish real-time monitoring system for dangerous driving behavior.Research shows that the risk of driving behavior identification based on hidden Markov model can accurately identify the model of dangerous driving behavior under conditions of highway snow and ice, but also for the proposed improvement measures to enhance highway safety under snow and ice conditions makes certain contribution.更多还原