基于多信息检测的车辆智能防撞预警技术研究  

Study on Intelligent Vehicle Anti-collision Forewarning Technology Based on Multi-information Detection

作　　者：刘志强[1] 汪澎[1] 秦洪懋[1] 仲晶晶[1] 宋世亮[1] 

机构地区：[1]江苏大学汽车与交通工程学院

出　　处：《中国安全科学学报》2010年第1期153-158,共6页China Safety Science Journal

基　　金：国家科技支撑计划课题(2007BAK35B02)

摘　　要：运用多源信息检测技术对驾驶状态进行综合检测,利用架设在车辆外部的摄像机获取本车与前方车辆之间的相对距离、相对速度信息,利用架设在车辆内部的摄像机获取驾驶人状态信息,通过综合判断降低车辆防撞检测的漏警率、误警率以及冗余报警率;构建车辆行驶安全判定规则,对行车实时危险和潜在危险进行检测、分析与判断,获取可靠的防撞警示提醒;使反应时间、相对距离、相对速度3个方面都能得到优化控制,从而避免车辆碰撞事故;同时也为提高车速、增加道路通行能力、实现自动驾驶等奠定良好的基础。Driving environment is very complicated, because of the multi-source information of the road, the relative distance with other vehicles or obstacles and driver＇s state, etc. To obtain more information about the driving, the state messages of the lane, relative distance and speed between the surrounding targets is identified by receiving information from the CCD camera in front of the car, and the driver＇ s attention state information is identified by receiving information from the CCD camera inside, which could effectively reduce the leak detection alarm rate, the false alarm rate and alarm rate of redundancy by using the technology of multi-source information detection. The real-time detection of dangerous or potentially dangerous driving information is used to analyze and judge the driver＇ s state through a constructed safety judg- ment rule, so as to acquire a reliable anti-collision warning, which is helpful to prevent vehicle collision and optimize the control of the relative distance, relative speed and reaction time. Meanwhile, it also sets a good foundation for the speeding, the increasing of traffic capacity and the realizing of automatic drive.

关 键 词：智能车辆 防撞预警技术 机器视觉 多信息检测 安全机制 

分 类 号：X924.3[环境科学与工程—安全科学]