杭明升

入学时间：1999年

答辩时间：2002年

论文题目：城市道路交叉口群实时自适应控制若干理论与方法研究

中文摘要

摘 要

本文主要对交叉口群实时自适应控制中的参数优化及相关的关键问题研究提出了相应的理论与方法。城市交通流从总体上讲，是一种“随机的局部、确定的总体”的间断变化流，因此，与工业中的实时自适应控制相比，显然，城市交通实时自适应控制的复杂性要高的多。对城市交通流进行实时自适应控制进行研究时，必须采取由控制机理解剖层→参数模型表达层→算法搜索求解层的方法，而不是在解析整个的控制机理之前就急于去确定控制参数的如何优化计算。在剖析国外相关系统的优缺点的基础上，结合中国城市道路交通存在的固有问题，以国际上迅速发展的智能交通系统为背景，提出了交叉口群实时自适应控制的概念和基本类型。然后，分别以混合交通流单点实时自适应控制、相邻两交叉口协调控制、相邻三交叉口协调控制为依托，使用解藕控制法和和藕建模法，分别对开放式线控条件下的信号周期、绿信比，相位差、相位相序优化提出了优化的目标函数、优化间隔和优化的算法。同时，对交叉口群协调控制的判定、短连线情况下的交叉口群协调控制、协调控制下的相位相序选择与优化、交通关联度影响下的战略参数优化及闭和网控下的相位差优化提出了解决方案。最后，对交叉口群协调控制中的过饱和情况下交通诱导与控制整合问题、优化算法实时自适应问题、协调控制的升降级问题、机动车LPU的确定、车流离散模型问题、检测器埋设的基本原则等相关的问题进行阐述。

关键词：交叉口群，实时自适应，城市交通控制，信号周期，绿信比，相位差，相位相序

英文摘要

Abstract

This research presents the theories and methods on optimizing parameters for grouped nodes under real-time self-adaptive urban traffic control. In general, urban traffic flow is an interrupted flow with “stochastic in locality, deterministic in overall”. So the complexity of urban traffic control is much higher than that of industrial self-adaptive control. When studying the urban traffic control, it must be in the order of : control philosophy layer→parameters model expression layer→algorithm searching layers, not in the contrary. Based on the booming intelligent transportation system and distinguishing traffic conundrum in china, the definition and types of grouped nodes are put forward after analyzing the strengths and shortages of traffic control system such as TRANSYT, SCATS, SCOOT. Taking isolated real-time self-adaptive, two neighboring intersections coordinated control, three neighboring intersections coordinated control as examples, the objection function, searching algorithm, refreshing interval of optimizing parameters(cycle, offset, split) under open-loop artery coordinated control are put forward by the decoupling control and coupled modeling. Under closed-loop grid coordinated control, Kruskal algorithm aiming at minimal costing tree is applied to eliminate the disturbs between two closed-loops when optimizing offsets with multi-closed-loops. At last, pertinent topics such as, coordinated control under over saturation integrating with traffic information guidance, real-time self-adaptive algorithm, how to calibrate LPU and models of platoon dispersion under mixed traffic, how to put loop detector in place, are briefly illuminated.

Keywords: Grouped nodes, Real-time self-adaptive, Urban traffic control, Cycle, Offset, Split, Phase Sequence