基于车联网的交通信息采集与事故识别方法研究
发布时间：2018-07-13 20:02
【摘要】：实时有效的交通信息采集能为交通诱导、交通拥堵、事故检测等提供高质量的数据,是城市交通规划和交通控制与管理的重要基础。目前交通信息主要通过线圈检测器、微波检测器、视频检测器等固定式检测设备进行采集,存在安装维护成本高、检测范围有限,难以提供全路网交通信息等不足,已无法满足智能交通系统对交通信息的需求。RFID技术凭借其多目标快速识别、抗干扰能力强、易部署、成本较低、寿命长、可携带车主和车辆信息等优势,在智能交通领域的应用越来越广,是交通信息采集技术发展的新方向。通过安装在路侧的RFID阅读器与车载电子标签进行通信,能够将所有车辆连入网络,实现车联网,可获得车牌号、车主信息等静态交通信息,这是其他采集技术所不具备的优势。因此,从理论和技术层面对基于RFID的车联网进行全面、深入地研究,具有重要的理论意义和实用价值。 本文对如何利用RFID技术构建车联网,以及利用车联网进行交通信息采集与交通事故自动识别进行研究,具体研究内容包括： (1)将RFID技术应用于智能交通系统中,并利用RFID技术构建车联网模型,针对车联网的应用特点,选取合适的RFID系统参数,为车联网的研究提供充分的理论基础。 (2)研究了车联网应用过程中的防碰撞问题,并对基于二进制树和基于ALOHA的这两类标签防碰撞算法进行仿真分析,结合车联网的实际应用特点,提出了一种适合于车联网的标签防碰撞算法,并利用MATLAB模拟采集过程,验证了本文防碰撞算法的效果,为解决车联网应用过程中的碰撞问题提供良好的底层算法。 (3)给出了交通量、密度、速度等传统可测参数和空间占有率、道路行程时间、道路延误时间、道路选择概率等车联网可测参数的检测模型和方法,并利用VISSIM仿真软件模拟真实的路网环境,对车联网的交通信息采集过程进行模拟,结果表明基于车联网的交通信息采集效果理想。 (4)研究了基于车联网的交通事故识别方法,通过分析事故发生前后交通流参数的特性,并根据车联网模型所采集到的交通数据特点,基于排队论思想提出了一种基于车联网数据的交通事故自动识别方法,并仿真模拟路网交通事故的发生,对本文算法的性能进行分析。实验证明,本文算法对事故的检测效率高,效果良好,且能够适应不同的道路交通状况。
[Abstract]:Real-time and effective traffic information collection can provide high quality data for traffic guidance, traffic congestion, accident detection and so on. It is an important foundation of urban traffic planning and traffic control and management. At present, traffic information is mainly collected through fixed detection equipment such as coil detector, microwave detector, video detector, etc., which has the disadvantages of high cost of installation and maintenance, limited range of detection, difficulty in providing traffic information of the whole road network, etc. RFID technology can not meet the demand of intelligent transportation system for traffic information. RFID technology has the advantages of fast multi-target identification, strong anti-jamming ability, easy deployment, low cost, long life, and can carry vehicle information, etc. It is more and more widely used in the field of intelligent transportation, which is a new direction of traffic information collection technology. Through the communication between the RFID reader installed on the road side and the electronic tag on the vehicle, all the vehicles can be connected to the network, and the vehicle can be connected to the network, and the static traffic information such as license plate number and owner information can be obtained, which is an advantage that other acquisition technologies do not have. Therefore, it is of great theoretical significance and practical value to conduct a comprehensive and in-depth study on RFID based vehicle networking from the theoretical and technical aspects. This paper studies how to use RFID technology to construct vehicle network and how to use it to collect traffic information and identify traffic accidents automatically. The specific research contents include: (1) applying RFID technology to intelligent transportation system. According to the application characteristics of vehicle networking, the appropriate RFID system parameters are selected to provide a sufficient theoretical basis for the research of vehicle networking. (2) the anti-collision problem in the application process of vehicle networking is studied. The two anti-collision algorithms based on binary tree and Aloha are simulated and analyzed. According to the practical application characteristics of vehicle networking, a label anti-collision algorithm suitable for vehicle networking is proposed, and the acquisition process is simulated by MATLAB. The effectiveness of the anti-collision algorithm in this paper is verified, which provides a good bottom layer algorithm for solving the collision problem in the application of vehicle networking. (3) the traditional measurable parameters such as traffic volume, density, speed and space occupancy are given. The detection model and method of the measurable parameters of vehicle network, such as road travel time, road delay time, road selection probability, etc. The real road network environment is simulated by using Visual IM simulation software, and the traffic information collection process of vehicle networking is simulated. The results show that the traffic information collection based on the vehicle network is effective. (4) the traffic accident identification method based on the vehicle network is studied, and the characteristics of the traffic flow parameters before and after the accident are analyzed. According to the characteristics of traffic data collected by vehicle networking model, a method of automatic identification of traffic accidents based on vehicle network data is proposed based on queuing theory, and the occurrence of road network traffic accidents is simulated. The performance of this algorithm is analyzed. Experimental results show that the proposed algorithm is efficient and effective in accident detection and can adapt to different road traffic conditions.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495
本文编号：2120582
[Abstract]:Real-time and effective traffic information collection can provide high quality data for traffic guidance, traffic congestion, accident detection and so on. It is an important foundation of urban traffic planning and traffic control and management. At present, traffic information is mainly collected through fixed detection equipment such as coil detector, microwave detector, video detector, etc., which has the disadvantages of high cost of installation and maintenance, limited range of detection, difficulty in providing traffic information of the whole road network, etc. RFID technology can not meet the demand of intelligent transportation system for traffic information. RFID technology has the advantages of fast multi-target identification, strong anti-jamming ability, easy deployment, low cost, long life, and can carry vehicle information, etc. It is more and more widely used in the field of intelligent transportation, which is a new direction of traffic information collection technology. Through the communication between the RFID reader installed on the road side and the electronic tag on the vehicle, all the vehicles can be connected to the network, and the vehicle can be connected to the network, and the static traffic information such as license plate number and owner information can be obtained, which is an advantage that other acquisition technologies do not have. Therefore, it is of great theoretical significance and practical value to conduct a comprehensive and in-depth study on RFID based vehicle networking from the theoretical and technical aspects. This paper studies how to use RFID technology to construct vehicle network and how to use it to collect traffic information and identify traffic accidents automatically. The specific research contents include: (1) applying RFID technology to intelligent transportation system. According to the application characteristics of vehicle networking, the appropriate RFID system parameters are selected to provide a sufficient theoretical basis for the research of vehicle networking. (2) the anti-collision problem in the application process of vehicle networking is studied. The two anti-collision algorithms based on binary tree and Aloha are simulated and analyzed. According to the practical application characteristics of vehicle networking, a label anti-collision algorithm suitable for vehicle networking is proposed, and the acquisition process is simulated by MATLAB. The effectiveness of the anti-collision algorithm in this paper is verified, which provides a good bottom layer algorithm for solving the collision problem in the application of vehicle networking. (3) the traditional measurable parameters such as traffic volume, density, speed and space occupancy are given. The detection model and method of the measurable parameters of vehicle network, such as road travel time, road delay time, road selection probability, etc. The real road network environment is simulated by using Visual IM simulation software, and the traffic information collection process of vehicle networking is simulated. The results show that the traffic information collection based on the vehicle network is effective. (4) the traffic accident identification method based on the vehicle network is studied, and the characteristics of the traffic flow parameters before and after the accident are analyzed. According to the characteristics of traffic data collected by vehicle networking model, a method of automatic identification of traffic accidents based on vehicle network data is proposed based on queuing theory, and the occurrence of road network traffic accidents is simulated. The performance of this algorithm is analyzed. Experimental results show that the proposed algorithm is efficient and effective in accident detection and can adapt to different road traffic conditions.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495
【参考文献】
相关期刊论文 前2条
1 马杨;;车路协同,还有多远?[J];中国交通信息化;2011年09期
2 刘建华;杨士航;;浅谈车联网技术发展与应用前景[J];中国高新技术企业;2010年28期
，本文编号：2120582
本文链接：https://www.wllwen.com/kejilunwen/jiaotonggongchenglunwen/2120582.html
