【作者】 孔祥杰；

【导师】 孙优贤； 沈国江；

【作者基本信息】 浙江大学 ， 控制科学与工程， 2009， 博士

【摘要】 城市路网交通流控制技术是近年来国内外控制领域和交通工程领域研究的热点之一。由于城市车辆的增长和路网密度的增加,交叉口之间的相关性逐渐增强,将城市路网中的多个交叉口作为一个整体进行有效的信号协调控制,以提高整个控制区域内的交通通行效率,已成为城市交通控制新的发展方向。随着自动控制技术、计算机技术、通讯技术和交通工程技术的迅速发展,各种交通流模型和协调控制算法层出不穷,新的理论和技术不断出现,有些已在实际交通工程中得到了成功的应用。本文讨论城市路网交通流协调控制技术。考虑到城市交通系统是一个具有严重非线性、随机性、时变性和不确定性的复杂系统,利用一次指数平滑预测方法、神经网络预测方法和模糊逻辑建立了短时交通流组合预测方法,并详细阐述了组合预测方法的结构和参数调整机制。同时,利用神经网络、模糊逻辑、遗传算法和大系统理论,设计了一些城市路网交通流智能控制算法,仿真分析和实际应用表明,这些算法具有较强的鲁棒性、自适应性和自学习性,与传统的交通控制方法相比,能更有效的解决城市交通问题。本文的主要内容如下:1.介绍了城市道路交通控制的起源和发展历史,详细的分析和阐述了国内外当前的研究成果,指出了理论研究和实际应用中存在的困难和一些亟需解决的问题,同时结合我国城市交通的具体特点和存在的问题提出了我国城市道路交通控制技术今后的研究方向。2.提出了一种短时交通流智能组合预测方法。该智能组合预测方法包括三个子模块:历史平均模块、神经网络模块和模糊组合模块。历史平均模块具有良好的静态稳定特性,神经网络模块对动态交通流量的预测具有较高的精度。为了充分利用上述两个单项模块对不同交通状况的适应性,采用模糊逻辑来综合这两个单项模块的输出,并把模糊组合模块的输出作为整个智能组合方法的最终交通流量预测值。3.设计了一种交通干线动态双向绿波带智能控制算法。整个控制方案分为两层,协调层根据一段时间内交通流数据计算公共周期时间、上下行相位差和绿信比,控制层实时调整各交叉口的绿信比并实现对信号灯的控制。周期依照关键路口饱和度的大小由模糊控制算法进行优化,而相位差根据上下行速度进行计算,绿信比基于历史和实时的交通数据确定。目标是使干线双向车流尽可能不停车的通过交叉口,明显降低车辆停车率和平均延误时间。4.对城市区域交通分布式协调控制进行研究,每个交叉口设置一个模糊信号控制器。控制器包括相位选择模块、绿灯观测模块和决策模块3个模块。相邻交叉口间的控制器相互协调,能够优化相位顺序和相位长。并用遗传算法训练模糊逻辑,提高了系统的鲁棒性。5.在分布式交通控制结构以及模糊理论和神经网络的基础上,提出一种具有公交优先的城市路网交通流协调控制算法。把整个路网作为一个大系统,路网中的各个路口作为子系统,每个路口设置一个智能信号控制器,核心控制部分由3个模糊决策模块组成,并用神经网络来实现模糊关系,提高系统的鲁棒性。目标是通过相邻路口控制器的信息交换和协调,实现整个路网交通流的协调和公交优先通行。6.详细讨论了绍兴市城区具有公交优先的路网交通流协调控制系统的实际方案。目标是实现绍兴市区相关路口交通信号的协调和公交优先通行,从整体上提高通行能力,减少车辆通行时间和路口停车率。在分析绍兴市交通现状的基础上,给出了具体的系统设计方案。方案中详细的提出了系统结构、控制方案和系统的软硬件需求等。并对绍兴市实施具有公交优先的路网交通流协调控制系统带来的预期效果和评价标准进行了阐述。最后对本文工作进行了概括性总结,并对进一步的研究做了展望。更多还原

【Abstract】 Control of urban traffic networks is a popular topic in control domain and traffic engineering domain at home and abroad in these decades. With the increase of the number of vehicles and the density of traffic networks, mutual influence of traffic flows among adjacent intersections is gradually strong. To increase efficiency of urban transportation, new urban traffic control technique regards the whole traffic networks as a large scale system and coordinatively controls every intersection at the same time. With the fast development of automatic control, computer technique, communication technique and traffic engineering technique, many traffic flow models and coordination control methods have been found. New theory and research achievements have been published in recent years. Some applications in engineering have shown their tremendous powers.In this dissertation we mainly study and analyze coordination control technique of urban traffic networks. We make use of intelligent hybrid forecasting method to forecast urban short-term traffic flow which is heavily nonlinear, stochastic, time-variant and uncertain. Moreover we describe the structure of intelligent hybrid forecasting method. We also design some advanced intelligent traffic control algorithms by use of neural network, fuzzy logic, genetic algorithm and large scale system theory. Simulation analysis and application results show that these algorithms are more robust, self-adaptive and self-learning, and can solve urban traffic problems more effectively than conventional traffic control methods.The main work and contributions of this dissertation are as follows:1. We make an overview on the generation, development and last achievements of urban traffic control technique at home and abroad in detail, and make a discussion on difficulties and problems in theory analysis and actual application. Combined with specific characteristics of urban traffic in our country, we present the future research direction on urban traffic control technique.2. In order to transcend the limitation of existing single forecasting technique on different traffic condition, a novel intelligent hybrid (IH) method for short-term traffic flow forecasting is presented. The IH method has 3 sub-modules: history mean (HM) module, artificial neural network (ANN) module and fuzzy combination (FC) module. The HM method has good static stabilization character. The ANN method can estimate the dynamic traffic flow in a very precise and satisfactory sense. In order to take advantage of the useful information of the HM module and the ANN module to improve the forecasting effect further, the two individual modules reflecting practical problems from different respects are combined by fuzzy logic. The FC module mixes the two individual forecast results and its output is regarded as the final forecasting of the traffic flow.3. A dynamic two-direction green wave intelligent control strategy is presented. The whole control structure is divided into the coordination layer and the control layer. Public cycle time, up-run offset, down-run offset and splits on the arterial are calculated in the coordination layer, and the splits of each intersection on the arterial are adjusted in the control layer at the end of each cycle. Public cycle time is adjusted by fuzzy logic according to the saturation degree of key intersection on the arterial. The offsets are calculated by average speeds. The variable splits of each intersection are adjusted based on historical and real-time traffic information. The target is to decrease vehicle average delay time and make vehicle stop as little as possible.4. An intelligent coordination control method of urban region traffic is presented. A fuzzy signal controller, including phase choosing module, green observation module and decision module, is installed at each intersection. Fuzzy signal controllers cooperating with each other can optimize phase sequence and phase length. In order to make the system robust, the fuzzy rules are optimized by genetic algorithm.5. On the basis of distributed road traffic control framework, fuzzy theory and artificial neural networks technique, an intelligent coordination control technique of traffic networks with bus priority is proposed. The whole traffic network is regarded as a large scale system and the subsystems are the intersections. Multi-phases intelligent signal controller that controls its own traffic and cooperates with its neighbors is installed at each intersection. The hard core of signal controller is composed by 3 fuzzy modules. In order to improve control system’s robusticity, the fuzzy relation of each module is implemented by a neural network respectively. The target of this proposed method is that through exchanging information from its own traffic detectors and its neighbors and cooperating among adjacent signal controller, social vehicle coordination and bus priority in whole traffic network are realized.6. We design intelligent coordination control system of traffic network with bus priority in Shaoxing City. The target is realizing social vehicle coordination and bus priority in traffic network of Shaoxing, increasing traffic capability, decreasing vehicle travel time and delay time. On the basis of analyzing the traffic situation of Shaoxing, we present the concrete design plan, which describes system structure, control plan, software and hardware design in detail. Moreover, we introduce the functions of the related modules. Finally, expected effects and evaluation criterion of intelligent coordination control system of traffic network with bus priority in Shaoxing City is discussed.Finally, we make a conclusion on current work and propose the future research directions.更多还原