基于时空检测的拥堵交通网络本征提取及快速疏散算法研究

Urban traffic is increasingly becoming a universal problem, which seriously affects the normal operation of road network and urban sustainable development. In order to alleviate the pressure on traffic management, the problem of rapid and effective traffic congestion evacuation needs to be solved urgently. Evacuation strategy must be on the basis of effective evacuation algorithm, therefore according to the temporal and spatial detection data of fixed detectors and floating cars, the research on rapid evacuation algorithm of traffic flow under the state of heavy urban road network has been carried out. Firstly, combined with the basic parameters of road network, the model for selecting boundary of congestion area, congestion evacuation affected area and coherent affected area has been established. Secondly, based on the defined affected area of congestion network and the dynamic OD data of residents travel, trend analysis model of traffic congestion area has been established on the basis of the dispersing rate. Thirdly, the model of route guidance in each crowded degree interval and driver path choice behavior disparity in the congestion area has been established by means of principal component analysis theory, and furthermore, under the collaborative optimization effect of macro dynamic traffic induction and traffic control, the rapid evacuation algorithm in traffic congestion area has been put forward. Fourthly, by selecting a typical network with perfect traffic control and management facilities, in-transit experiment and parameter calibration have been carried out in order to make the above algorithm be able to adapt the actual characteristics and provide scientific theory and method for promoting traffic operation efficiency of congestion roads and improving traffic condition .

拥堵日益成为城市交通的普遍问题，严重影响道路网络正常运行和城市可持续发展，为了缓解交通管理压力，快速、有效的交通拥堵疏散亟待解决。疏散策略必须依据有效的疏散算法，因此基于固定检测器和浮动车的时空检测数据，针对拥堵发生状态下城市道路网络交通流快速疏散算法进行研究。首先结合道路网络基本参数，建立拥堵区域、拥堵疏散影响区和相干影响区边界选取模型；其次以界定的拥堵路网影响区和居民出行动态OD数据为基础，建立基于离去率的交通拥堵区变化趋势分析模型；再次采用主成份分析理论建立拥堵区域各拥挤度区间路线指引与驾驶员路径选择行为差异模型，进一步提出宏观动态交通诱导与交通控制协同优化作用下拥堵区交通流快速疏散算法；最后选取设置有完善交通控制与管理设施的典型路网，进行在途实验及参数标定，使建立的拥堵交通流快速疏散算法更能适应实际特征，为提升拥堵道路交通运行效益、改善交通拥堵状况提供科学的理论与方法支持。

为了缓解交通拥堵，掌握交通拥堵的变化态势并实施快速有效的交通诱导与控制策略亟待解决。根据数据—模型—实验的研究思路，项目主要研究工作有：（1）分析了多源交通信息检测数据提取方法，根据交叉口交通流的运动变化规律，将日交通流变化划分为高峰时段、低峰时段和平峰时段三个阶段，通过均衡地磁式无线车辆检测器网络的能量负载，提出了无线通信分簇算法与路由协议。（2）根据浮动车和固定检测器检测拥时空数据，通过分析拥堵路径交叉口关联度建立了交通网络瓶颈划分模型。（3）根据瓶颈发生类型，以路网历史和实时交通量为参考对象，建立了基于自回归分析的拥堵瓶颈交通需求预测模型。通过比对预测交通量与拥堵路网各个节点实时流出能力，建立了拥堵区拥堵状态及拥堵时间决策模型。（4）在分析路网状态参数对宏观基本图的影响基础上，通过最小生成树的初步划分、基于归一化分割的深度划分和反向合并过程，提出了MST-Ncut的MFD （macroscopic fundamental diagram）子区组合划分方法。（5）通过分析MFD子区之间的车辆流入流出关系，建立了宏观网络车流平衡方程，进而以路网旅行车辆完成率最高同时子区边界处受阻车辆数最低为优化目标建立了拥堵区边界最优控制方法。（6）为了缓解宏观交通网络拥堵区的拥堵状况并降低机动车燃油消耗，以宏观路网出行车辆完成率最高同时能耗最低建立了能耗节约拥堵区边界双目标优化控制和Fuel Consumption-Proportional Integral 控制方法。（7）通过问卷调查，分析了VMS信息环境下驾驶员路径选择行为规律，以此为基础通过协调交通诱导与控制策略，提出了拥堵区快速疏散算法。（8）基于地面路网MFD，以快速路出口匝道通行能力和与之相衔接的地面路网承载能力为约束条件，以整个路网的车辆总行程时间最短为优化目标，建立了快速路出口匝道流量分配模型。采用实测或仿真数据验证表明，项目提出的相关控制方法能够有效的防止子区进入拥堵状态，提高了路网旅行车辆完成率，可为提升拥堵道路交通运行效益、改善交通拥堵状况提供科学的理论与方法支持。