【作者】 郭伟；

【导师】 杨明；

【作者基本信息】 上海交通大学 ， 控制理论与控制工程， 2012， 硕士

【摘要】 车路协同作为智能交通运输系统的重要组成部分,为缓解道路拥挤、提高道路通行能力、改善道路交通安全等发挥了重要作用。采用实际智能车进行车路协同研究成本过高,然而软件仿真平台的实验环境过于理想,因此,本文设计并实现了车路协同硬件仿真系统。本文所设计车路协同硬件仿真系统由仿真车系统和道路设施系统组成。针对视觉导航可靠性较差的问题,本文仿真车采用电磁导航方式,仿真车采用里程计结合RFID射频识别特征路标的方式实现车辆的定位,采用WiFi无线传输技术实现车辆间的无线通信。实际实验表明,该仿真车能够在动力学、导航、定位、无线通信等方面很好地仿真实际智能车。本文设计了正弦信号发生器用于产生一定频率的交流电源驱动赛道下方的导引线。通过比较目前信号发生器的不同设计方案,设计了基于精密函数发生芯片MAX038的信号发生器。该信号发生器主要包括信号生成模块、输出放大模块、电流和频率测量模块、人机交互模块。为了保证导引信号的频率稳定性,设计了基于单片机的频率闭环控制电路。该信号发生器通过闭环控制保证了输出信号的频率稳定性,实验结果表明该信号发生器完成了设计目标,具有一定的稳定性和可靠性。高效的路口多车协作可以减少道路拥堵的情况发生,提高交通流量和交通系统运行的效率。针对无交通灯十字路口,本文提出了一种基于博弈论的路口多智能车协作算法。首先,利用场图工具建立路口车辆冲突模型。然后,针对冲突车辆设计了基于博弈论的协调算法。使用本文开发的硬件仿真平台,对提出的算法进行了验证。实验结果表明,采用本文算法车辆通过路口的时间比采用冲突表算法缩短了约40%,本文提出的算法在保证车辆安全通过路口的同时,提高了路口的通行效率。本文针对双向双车道交通场景下的智能车辆超车问题进行了研究,提出了基于反应式智能体的超车算法。在车辆超车过程中,根据车辆对周围环境的感知建立相应的虚拟环境,在虚拟环境中建立反应式多智能体系统,根据物理学启发定义智能体之间以及智能体与虚拟环境间的相互作用,最后通过对多智能体的空间分布的全局评价指标得到车辆超车动作的控制指令。经过实验表明,本章提出的算法可以保证车辆超车过程中的安全性。更多还原

【Abstract】 Vehicle-to-Infrastructure cooperation system has become the forefront of technology and research focus in the field of intelligent transport system. It plays an important role in easing the road congestion, increasing the road capacity and improving the traffic safety. The cost is too high in the research of cooperative vehicle infrastructure system using actual intelligent vehicle. However, the experimental environment is ideal in software simulation. For these reason, this thesis designs and implements a hardware simulation system for vehicle-to-infrastructure cooperation system.The hardware simulation system designed in this thesis consists of mini intelligent vehicle system and road infrastructure system. The mini intelligent vehicle use electromagnetic navigation, owing to the poor reliability for visual navigation. The mini intelligent vehicle implements its longitudinal localization through combination of odometer and road sign detection using RFID,and the communication between vehicles is based on WiFi wireless transmission technology. In summary, The mini intelligent vehicle can simulate real intelligent vehicle in kinetics, navigation, localization and wireless communication.By comparing the different design solution of signal generator, a sine signal generator based on MAX038 is designed to produce a certain frequency AC power to drive the guide line. The basic research content includes sine signal generator and current amplification module, current and frequency measure module, human-computer interaction module. The signal generator provides close-loop stable control of frequency, to make the frequency output smoothly. The experiment revealed that the signal generator could generate a sine signal with high precision and steady that meet with theoretical design.Intelligent vehicles’cooperative driving in intersection can reduce the congestion occurs and improve the passing rates. The thesis introduces an algorithm based on dynamic game for intelligent vehicles cooperation at the unsignalized intersection. Conflict between intelligent vehicles is described by using field graph tool. A conflict avoidance algorithm was built up with dynamic game theory. The hardware simulation system designed in this thesis is used to test the algorithm. The algorithm not only coordinates the intelligent vehicles among intersection without collision,but also improve the passing rates by about 40% compared to the collision table based algorithm. The thesis presents an algorithm based on reactive multi-agent system to the overtaking problem of vehicle. During the overtaking process, the overtaking vehicle creates a virtual environment based on vehicle’s perceptions of the material environment around it. A decision vector for the vehicle is elaborated from the evaluation of a set of indicators characterizing the global state of a system of reactive agents (RMAS), which is evolve in the virtual environment. Agent-to-agent and agent-to-environment interactions are inspired by physics, in order to produce a distribution of agents over the virtual environment. Experiments show that the algorithm can guarantee the safety of vehicle during the overtaking process.更多还原