【作者】 钱喆；

【导师】 徐建闽；

【作者基本信息】 华南理工大学 ， 交通信息工程与控制， 2014， 博士

【摘要】 随着城市经济的发展，机动车保有量呈现快速增长的态势，而城市道路增长有限且缓慢，导致城市的交通出行严重受阻，交通拥堵频发，交通过饱和状态已经成为常态。如何从整个交通系统的战略角度出发，采取有效的交通措施缓解过饱和交通问题已经成为城市发展的新要求。本文从交通信号控制入手，针对现有过饱和信号控制理论方法的不足，结合过饱和交通状态的特征，在过饱和交通状态识别、过饱和单点控制模型建立、过饱和干道协调控制模型建立以及过饱和交叉口群的协调控制等方面进行了相关的研究与探讨，主要的研究成果可概括如下：1.通过分析各种不同对象的过饱和交通状态，给出了过饱和交通状态的定义，进一步说明了在过饱和状态下，交叉口存在绿灯时间使用率高、进口道排队积累以及车辆延误显著增加的交通特性，利用过饱和下的排队长度较长特性，提出了基于最大排队长度的过饱和状态判别方法，该方法简单易行但存在一定的局限性，为了更有效地识别过饱和交通状态，利用检测器的检测数据建立了滞留排队长度预测模型和上游交叉口溢流预测模型，基于这两个模型提出了两个描述过饱和现象的两个交通参数（排队消散系数和溢流阻滞系数）的计算公式，并通过算例分析验证了这两个系数可以有效识别过饱和交通状态。2.分别论述了在过饱和交通状态下单交叉口的动态信号控制、排队长度相等、主动排队控制和延误最小化控制等四种不同控制策略，以延误最小化为主要控制目标，兼顾其它三种策略建立了延误连续控制模型和延误离散控制模型，延误连续控制模型以连续车流为对象，从进口道的延误计算入手构建了交叉口整体的延误计算模型，说明通过优化求解可得到交叉口的最优控制方案，而延误离散控制模型以单位时间（周期）的车流为对象，从交叉口的整体延误计算入手建立了交叉口的控制模型，设计了遗传算法和极值转换算法分别求取离散控制模型，以离散控制模型求解算例说明了模型的合理性。3.以过饱和干道单方向的通行能力最大为目标，提出了单向协调控制模型的目标函数，分析了模型的约束条件，提出了过饱和下的四种不同排队控制策略，基于这四种排队控制策略建立了过饱和下的单向协调控制模型，以单向模型的协调主方向为前提，说明了干道次方向存在交通阻滞、交通补偿和绿灯空放三种交通状态，进一步建立了不同交通状态下进口道放行量的计算公式，综合两个方向的约束条件，建立了以主方向优先的过饱和双向协调控制模型，通过算例分析了不同排队控制策略下干道的控制效果，并对比分析了四种策略的优劣，验证说明了模型的正确性。4.分析了过饱和状态下交叉干道的协调控制方法，提出了采用不同的优先级别实现交叉干道的协调控制，并设计了不同优先级别的干道权重系数的计算过程，介绍了交叉干道协调控制子区的划分方法和子区间的路段集合，同时对这些协调控制基本单元进行了参数定义，重点说明了子区间路段集合的权重计算公式，设计了过饱和下的交叉干道协调控制的步聚，方法从权重最大的子区入手基于路段权重大小逐步将各个子区合并直至将整个区域合并，基于协调控制步骤建立了交叉干道协调控制模型，通过算例进一步说明了过饱和下交叉干道的协调控制实现过程。5.分析说明了群体动力学理论与方法，提出采用递阶协调的方式、动态调整和优化交叉口的周期和绿信比实现过饱和下的交叉口群协调控制方法，基于群体动力学原理，建立了交叉口绿信比优化调整的动力学算子，并对其稳定性做出了分析与说明，对过饱和下交叉口群的合并与分离条件进行了定义，并针对这两种情况设计了周期和绿信比等的调整算法，综合两种算法提出了过饱和下的整体动态协调控制算法，同时以过饱和排队延误最小建立了过饱和下的交叉口群协调控制模型，说明其优化求解流程，最后以广州市中山路为案例说明模型的优化求解结果，并与MAXBAND方法、TRANSYT-7F方法的协调效果对比验证说明模型的优越性。6.以广州市天河区天河北路为案例，运用过饱和交通识别方法判定天河北路的过饱和交通状态，利用延误控制模型优化各交叉口的绿信比，运用干道协调控制模型优化干道整体的配时方案，并通过交通仿真分析天河北路在改善前、绿波协调控制下及改善后的延误时间及停车次数指标的变化情况，验证说明了论文研究成果的正确性与实用性。更多还原

【Abstract】 With the city ’s economic development, the vehicle in the city is growing rapid,but the growth of urban roads is limited and slow. It causes the traffic jams and thetravel time is increasing. The traffice oversaturation has been normal in the city. Sohow to take effective measures to alleviate the traffic problem from a system point ofview has become a new requirement for development. Considering the defects inexisting theory and method of oversaturation signal control, combining with thecharacteristics of oversaturation traffic, this dissertation has researched and discussedin the following areas, including the indentification of oversaturation traffic, a singleintersection control model in oversaturation, arterial coordination control model inoversaturation, the intersection group coordination control and so on. The mainscientific research and research results in this dissertation include:1. By analyzing a variety of different oversaturated objects, the definition ofoversaturation is given, and the traffic characteristics, such as high green time usagerate, approach queue accumulation and the increasing vehicle delay, are illustrated inoversaturation. Using the characterist of the long queue length, the identificationmethod of oversaturation based on the the maximum queue length is proposed, whichis easy to take out but there are some limitations. In order to more effectively identifyoversaturation, a retention queue length predictive model and the upstreamintersection overflow prediction model are establish using the detection data of thedetector, and based on these two models two traffic parameters(queue dissipationfactor and the overflow block coefficient) is calculated which describe theoversaturation status. By the numerical example these two factors are verified thatthey can effectively identify the traffic state of oversaturation.2. Four different control strategies in oversaturation are respectively discussed,which including traffic signal control dynamically for single intersection, equal queuelengths, active queue control, delay minimization control. Based on the minimum ofdelay, and taking the other three kinds of strategy into account, the continuous controldelays model and the discrete control model are established. For the continuous delaycontrol model which is based on a continuous traffic flow，the overall intersectiondelay model is built by the approach deley calculation, and the method to obtain theintersection optimal control scheme is indicated. For the discrete delay control model which is based on the unit time traffic flow, the intersection control model isestablished by the the overall delay calculation, the genetic algorithm and theconversion algorithm are designed to strike the discrete model. Finally, thereasonabilty of the discrete control model is illustrated through the numercialexample.3. The objective function of the one-way coordination control model is proposed,whose objective is the the arterial maximum capacity. Four different queuing controlstrategies in oversaturation is pointed out, and based on those the one-waycoordination control model is built. Under the premise of the one-way model, theother direction of the arterial has thress status, such as roads blocked, trafficcompensation and wasted green time. In the different traffic status, the calculations ofthe approach volume are established. So combining with the constraints in bothdirections, the bidirectional coordination control model in oversaturation isestablished, which is priority to the main direction in arterial. By the numericalexample, the control effect in different queueing strategies are analyzed, and the prosand cons of four strategies are also Comparatively analyzed, and those can verify thecorrectness of the model.4. The oversaturation coordination control method of the crossing arterial isanalyzed, the way of using different priority levels to achieve coordination in crossingarterial is proposed, and the calculation process of weight coefficient of the differentpriority level arterial is designed. The crossing arterial can divide to sub-area andsection set, so the coordination control sub-area division method and the section setbetween sub-area are introducted, and the parameter definitions of those coordinatedcontrol basic unit are introducted. The coordination control procedures is designed,which start from the biggest weigh sub-area and gradually merger other sub-areasuntil the entire region merging based on the weigh of section. Then based thecoordination control steps, the crossing arterial coordination control model isestablished, and the implementation process of the model is further illustrated by thenumerical example.5. The population dynamics theory and methods is analyzed, and thecoordination control method of intersection group is proposed by using hierarchicalcoordinated approach, dynamically adjust, optimize the intersection cycle and greensplit. Based on population dynamics, the optimal adjustment dynamics operator ofintersection green split is established, and whose stability is analyzed and explanated. The conditions of the oversaturation intersection group merger and separation aredefinited, the adjustment algorithms of the cycle and green split for these two casesare designed. Integrating these two algorithms, the overall dynamic oversaturationcoordination control algorithm is established. The intersection group coordinationcontrol model is built based on the queuing delay minimum, and whose optimizationsolving process is indicated. Finally, take the Guangzhou Zhongshan Road for examle,the optimal solution result of the model is illustrated, and compared with thecoordination effect of the MAXBAND and TRANSYT-7F, the superiority of themodel is described.6. Taking the Tian-He-Bei Road in Guangzhou Tianhe District as a case, theoversaturated traffic status of the Tian-He-Bei Road is identificated using theoversaturated traffic identification method, the split of all the intersections isoptimized by the delay control model and the arterial signal program is optimized bythe arterial coordination control model. So, using the traffic simulation, thecorrectness and practicality of the research is verified by analyzing the delay andnumber of stop in defferent states, such as before improvement, on green wavecoordination control and improvement.更多还原