【作者】 俄文娟；

【导师】 王云鹏；

【作者基本信息】 吉林大学 ， 交通环境与安全技术， 2012， 博士

【摘要】 道路交叉口行车安全问题一直是世界各国面临的共同难题，一直困扰着众多学者。因为交叉口是道路交通网的重要枢纽点，连接不同方向的道路，确保道路交通网中的车辆实现自由转向，但易引发不同车流的冲突，是道路交通事故多发地。目前，针对交叉口安全问题的研究主要集中在人、车、路、交通环境等交通影响因素及其与交通事故关系分析，并以此为依据改善道路设计、加强交通管控措施等，或者在车辆上安装主被动型安全装置以此来提高道路行车安全，但这些传统的简单的道路设计、交通控制和管理技术只能在一定程度上减少交通流的冲突数或降低冲突的严重程度，并不能有效地防止事故的发生，无法达到交叉口安全效益最优，更难以在本质上解决交通问题，尤其对于提高无信号交叉口安全水平更是无济于事。随着全球定位技术、无线通信技术和智能车辆技术的发展，车路协同技术逐渐成为解决交通安全问题的有效手段。车与车之间、车与路之间可以实现信息共享与交互，对交叉口的车车冲突状态进行预测和评估，通过冲突消解，避免碰撞事故的发生；车辆与交通信号灯进行通信，车辆可以获知绿灯的剩余时间，通过判断自车在剩余绿灯时间内是否能安全通过交叉口，防止车辆进入信号灯“困惑区”；在路段同向行驶中，后车可快速接收到前车紧急制动信息，进而避免追尾事故发生；在换道过程中，换道车辆通过感知周围车辆的状态，确定合适的换道时机，并通过无线通信，把换道信息传递给周围车辆，从而提高换道行驶安全性。目前，车路协同技术已经成为全球智能交通领域的研究热点，将引领智能交通的发展，对于大幅度提高道路交通安全具有重要意义。本文针对道路交叉口行车安全问题，依托国家863计划“车车交互式协同控制系统关键技术”（课题编号：2011AA110403），以无信号交叉口为研究对象，基于车路协同思想，设计无信号交叉口两车冲突检测方法与冲突消解策略，使车车冲突在向事故演化之前实现消解，消除因冲突造成的交通事故，为实现多车协同控制提供理论基础。研究工作如下：(1)基于信息交互的车车冲突识别方法研究。分析了全球定位系统的定位原理和GPS坐标转换过程，解析WGS-84大地坐标系到局部平面坐标系的转换方法；分析车与车之间需要获取的信息种类，包括车辆位置（经纬度）、速度、加速度、航向角以及车辆尺寸等，结合卡尔曼滤波方法和车辆运动学方程，建立了基于卡尔曼滤波的车辆位姿预估方程；搭建了车载信息采集系统，在校园内进行了实车验证，结果表明本文建立的车辆位姿预估方程的准确性；然后建立了考虑车辆外形尺寸的两车冲突检测方法，选取TTC作为冲突危险程度指标，确定了冲突严重程度级别。(2)无信号交叉口通行规则库建立。选取两车道十字形交叉口为研究对象，对交叉口区域车车冲突进行了分类，进而分析了由车车冲突导致的交叉口事故特征；以道路交通安全法通行规定为基础，面向交叉口车车冲突消解，研究不同冲突状态下车辆直行、左转、右转等过程的通行规则，建立基于车车通信的交叉口车辆通行规则库。(3)两车冲突消解策略研究。通过对两车冲突消解进行理论分析，指出两车冲突是两车在通过交叉口时在时间和空间上彼此之间有叠加的部分，若能适当调节车辆进入交叉口的时间，则会有效避免两车碰撞事故的发生；结合现有交叉口冲突算法的不足，分别针对合流冲突和交叉冲突，根据所建立的无信号交叉口通行规则，以车辆安全通行为优化目标，建立交叉口合流冲突消解模型和交叉冲突消解模型，通过优化不同车辆通过交叉口的时间与次序，实现交叉口车车冲突的消解；将冲突消解算法转换成Simulink逻辑图进行求解，为后续仿真提供基础。(4)交叉口多车协同冲突消解仿真。研究了基于信息交互的多车协同问题，建立了交叉口多车协同冲突消解决策流程；使用PreScan软件搭建了多车协同仿真实验系统，包括交叉口交通环境和车辆模型的搭建；设计仿真方案，包括两车冲突消解、三车冲突消解和四车冲突消解，对车车冲突检测方法和冲突消解模型进行仿真验证，将优先级别最高的车辆运行状态作为输入，通过冲突消解模型的计算，优先级别较低的车辆均可获得本身的调速方案，且实现不停车通过交叉口，仿真结果表明了提出的车车冲突消解算法的有效性；分析了关键参数（车辆外形尺寸、车辆定位误差、信息传输迟滞等）对冲突消解策略的影响，给出了误差校正方法和冲突消解修正模型，为进一步提高冲突消解策略的可靠性奠定了基础。更多还原

【Abstract】 Traffic safety on the road intersection has been a common challenge in various countries,and the problems of intersections has troubled many scholars. Because the intersectionscomposed by different directions of the road is an important traffic hub in a road net, and manyvehicles steer freely at the hub, which is easy to produce different traffic conflicts. So the traffichub is accident-prone area. Now, the research beased on intersection safety is focused on thetraffic impact factors, such as people, vehicle, road and transportation environment and theirrelationships with traffic accident. In order to improve the traffic safety, researchers use theanalysis results to perfect the road design, improve the traffic control measures, or install theactive/passive safety devices on the vehicles. But the traditional design, control andmanagement technology can only take effect on some problems in a certain extent, such asreducing the number and severity of traffic flow conflict, which couldn’t prevent accidents,achieve the optimal safety profits, or even solve the traffic problems in essence effectively,especially have little effect to improve traffic safety at unsignalized intersections.With the technical development of the GPS, wireless communication and intelligentvehicle, VII (Vehicle Infrastructure Integration) has been becoming an effective means to solvethe traffic problem. Under the VII environment, information sharing and interactivity can beachieved via V2V (Vehicle to Vehicle communication) and V2I (Vehicle to Infrastructurecommunication). Multi-vehicles cooperation can accomplish risk detection and evaluation,avoiding the traffic accidents. For example, the vehicles at intersection can communicate withthe traffic lights via information sharing and interactivity can be realized via V2V and V2I, toget acknowledge whether can it passes the green-signalized intersection safely, and to get rid ofdilemma zone; The leading vehicle conveys its starting information to the following vehicle，toinduce the waiting time and increase the intersection capacity. Between the running vehicles tothe same direction, the leading vehicle conveys its braking information to the following vehicle,avoiding the rear-end accidents. In the changing line process, the vehicle determines the propertiming to change line by sensing the status of vehicles around, to increase the driving safety.Nowadays, VII has been the research focusing on global intelligent traffic field, leading thedevelopment of the intelligent traffic, and greatly improving the traffic safety.This dissertation, which is sponsored by the National High Technology Research andDevelopment Program (863) (No. 2011AA110403), focus on the traffic safety peoblem on road intersection. On the basis of VII, two vehicles conflict detection and resolution strategies at anunsignalized intersection are innovatived, which eliminate conflicts before they turn intoaccidents. It provides a theoretical basis for the realization of multi-vehicle cooperative control.The major research work includes the following aspects:(1) Vehiceles conflict identification method based on vehicle to vehicle communicationis researched. At first, vehicle appropriate positioning method and technique of GPS coordinatetransformation are analyzed. Then the problem about the coordinate transformation fromWGS84 coordinate system to the plane coordinate system is discussed.The kinds of informationthat need to get between the vehicles is determined, including vehicle location (latitude andlongitude), speed, acceleration, heading angle, and vehicle size. Kalman filtering method andvehicle kinematic equations are analysed, and a vehicle trajectory prediction equation isestablished based on Kalman filter. After vehicle information collection system is builded, realvehicle experiment is carried out. The experiment results show that accuracy of the proposedvehicle trajectory prediction equation. Then a vehicle collision detection method consideringthe vehicle’s dimension is established, and TTC (Time to collision) is selected as an indicator ofthe degree of vehicle conflict risk.(2) Traffic rule of unsignalized intersection is designed. The vehicles conflicts at thecrossing are classified, and then the characteristics of intersection accidents caused by thevehicles conflicts are analyzed. Based on the Road Traffic Safety Law, passage rules of theintersection when the vehicle will go straight, turn left, turn right and so on are researched. Andthen passage rules of the unsignalized intersection are designed.(3) Vehicle conflicts resolution models are designed. Atheoretical analysis about conflictresolution of the two vehicles is carried out. It is pointed out that the vehicle conflict, in essence,is two vehicles which will pass through the intersection at the same time, which means that theyhave the superimposed part of time and space. If one of them appropriately regulates the timeof entering the intersection, it will effectively prevent two-vehicle collision accident. Afteranalyzing disadvantages of the existing intersection collision algorithm, the confluence conflictand cross-conflict resolution model are designed according to the unsignalized intersectionrules to make sure the veihicles at intersection driving safely. These models determine the timeand sequence of passing through the intersection for different vehicles.(4) Simulation experiments of multi-vehicle cooperative conflict resolution atunsignalized intersection are researched. Multi-vehicle coordination problem under vehicle tovehicle wireless communication environment is reseached, and multi-vehicle cooperativeconflict resolution decision-making process at unsignalized intersection is designed. Then amulti-vehicle coordination control simulation system is built by using PreScan software and the vehicles conflict detection and conflict resolution model simulation experiment is carried out.In the simulation experiment, three cases of multi-vehicle cooperative conflict resolution aredesigned. The vehicle which has low priority should adjust its velocity according to the statusof the vehicle which has high priority to pass the intersection. The simulation results show thatall of the vehilcles can pass through the intersection without collision, which means that motionof individual vehicles can be conducted in a safer, more deterministic and smooth mannerunder cooperative conflict resolution guidance. Simulation results indicate that the proposedconflict resolution algorithms have the potential for actual applications in driving at blindcrossing. In addition, the factors (Vehicle dimensions, vehicle positioning errors, informationtransmission delay, etc.) that affect the conflict resolution strategies are analyzed, and thenfurther amendments of the conflict resolution model are researched, which provides a basis forimproving the efficiency and reliability of the conflict resolution strategies.更多还原