基于虚拟供需闭合区域的受控智能路网拥挤控制理论和方法

As the intelligent transportation system constructions at home are increasingly perfect, the originally open network system of urban road traffic is gradually changed into a virtual closed controlled network, in which the traffic processes and states can be detected, measured and controlled. But the traditional open system idea based theories about the balance between supply and demand of urban traffic system, as well as the corresponding congestion control theories and methods, are difficult to match effectively with current field applications. How to identify urban traffic network equilibrium principle under the condition of intelligent transportation systems and set up the corresponding theory and method of congestion control becomes the crucial issues for modern urban traffic management and smart city construction. This project explores the matching relations among flow-in-and-out rates of different detection sections based on the detected information of multimode intelligent transportation system, traces and correlates the traffic flow origins and destinations as well as movement processes so as to create traffic state transition matrix based on Markov Chain and traffic flow coupling matrix of the controlled network, then traffic state prediction covering the congestion process is realized. This project also identifies closed perimeter or boundary expressed as the discrete detection sections of traffic flow-in-and-out data in the range of controlled urban road network so as to establish a virtual closed spatio-temporal system with deterministic, proposes traffic equilibrium principle of urban road network controlled by ITS, which describes the completely closed relationships among the different spatio-temporal data of traffic demand and capacity supply. In this research active control technology is introduced, thus an integrated model of traffic congestion network pinning control mixed with active control are constructed in combination with Macroscopic Fundamental Diagram (MFD) theory and the perimeter and boundary flow control methods, which take advantage of pinning control theory that is propitious to solve the spatio-temporal chaos of complex network. And relative empirical studies are conducted to ensure that the above results could provide theoretical support and appropriate technology for the practical intelligent transport system applications.

智能交通系统将开放性城市道路交通网络转变为过程和状态可测可控的虚拟封闭受控路网，而传统的基于开放系统理念的交通平衡原理和相应的拥挤控制理论及方法难以与现场应用有效对接。如何辨识智能交通系统条件下城市交通网络平衡原理并建立相应的拥挤控制理论及方法，已成为现代城市交通治理及智慧城市建设的关键问题。本项目基于多种类智能交通系统的检测信息，发掘受控路网上检测断面出入流率的匹配关系，追溯和关联受控路网范围内交通流起讫点及其运动过程,建立交通状态转移矩阵和交通流耦合矩阵，实现状态过程化预测；辨识闭合“数据边界”以构建具有确定性的虚拟封闭时空系统，建立完全供需闭合的网络交通流平衡模型；研究中将利用牵制控制理论在处理复杂网络时空混沌问题上的优势，引入主动式控制策略，结合宏观基本图理论和边界控制方法，构建交通拥挤主动式混合牵制控制集成模型，并进行实证研究和仿真验证，为智能交通应用提供理论支撑及适用技术。

源自开放系统理念的交通平衡原理和拥挤控制方法难以与现场应用有效对接，需深入发掘智能交通系统给传统开放性城市路网交通带来的时空虚拟封闭和状态可测可控的理论技术特性。本项目基于多种类智能交通系统的检测信息，发掘时空路网上交通拥挤传播规律与拥挤因果关系，追溯和关联受控路网范围内交通流起讫点及其运动过程,建立拥挤传播路径溯源和预测模型，从而实现城市路网状态过程化预测，以交通状态的能观化表达，为路网控制提供了完备、精准的数据支撑。基于宏观基本图理论，辨识控制区域的虚拟闭合边界并构建其理论模型，实现了对受控区域作为确定性虚拟封闭时空系统的准确描述，这扩展了路网状态能观化的内涵并为路网级别的控制奠定了基础。根据传统交通平衡模型和城市路网所展现的宏观特性，建立起了完全供需闭合条件下的网络交通流平衡模型，由此扩展了路网的可控性表达。本项目利用牵制控制理论和人工智能技术在处理复杂网络时空混沌问题上的优势，引入主动式控制策略，结合宏观基本图边界反馈控制方法和深度强化学习理论，建立了主动式交通拥挤异常发现与控制集成模型，并提供了计算机应用软件的实现方案。本项目理论研究结合仿真验证和现场实用进行，还开发完成了城市路网智能交通控制子区划分可视化交互处理和干道绿波协调控制区域划分及其方案生成的计算机应用系统软件，为城市路网智能交通控制提供了理论支撑及适用技术。