多服务公交系统协调控制理论与方法

A transit route usually adopts multiple service patterns, such as all-stop, short-turn and express services, to improve transport efficiency and level of service. If service patterns are not well coordinated, the stability and reliability of bus operations cannot be guaranteed, resulting in the low attraction of transit for the public. The most obvious phenomenon of unstable and unreliable transit operations is bus bunching. To alleviate the bunching problem, researchers proposed many control methods. However, most of them are suitable only for a bus route with a unique service pattern. This study focuses on real-time control for a bus system with multiple service patterns. Firstly, this study analyzes the factors that cause bus bunching and the formation process of bus bunching, and then develops a model to predict the probability that bus bunching will occur. Based on the idea of active control, this study develops a robust control method for bus operations and designs the control mechanism of "coordinate optimization, synergy control and feedback adjustment". That is, the proposed approach coordinates the operations of buses of multiple service patterns, realizes the cooperative operations of holding, signal priority and speed adjustment, and adjusts control strategies based on feedback information. Eventually, the transit system under control would lie in or approach to the stable condition. Research findings will reveal the law of the instability of transit operations and provide theoretical foundation and technical supports for transit control and management.

为提高运输效率和服务水平，公交线路常采用全程车、区间车和大站快车等多服务模式组合。如不同服务模式缺乏有效协调控制，公交运行的稳定性和可靠性无法得到保障，导致公交对公众的吸引力降低。公交运行不稳定和不可靠的最直观表现为公交串车。为缓解公交串车问题，国内外学者提出许多控制方法，但多数仅适用于单服务公交路线。本项目以多服务公交系统为研究对象，面向实时控制的需求，从公交串车的影响因素分析入手，解析串车的形成机理，建立"公交串车预测"模型以准确刻画串车的可能性。然后基于主动控制的思想，构建公交鲁棒控制方法和"协调优化、协同控制、反馈调节"的控制机制，协调多服务模式的公交运行，协同重点站驻站、信号优先和车速调节三类控制手段，根据反馈信息实时调整控制方案，保证系统运行处于或接近稳态。成果将揭示公交运行失稳的规律，为公交控制管理提供理论基础和技术支撑。

项目以多服务模式公交路线为研究对象，对数据获取、运行机理解析、优化设计和优化控制进行了系统研究。主要成果包括（1）建立了战略-战术层面协调的多模式公交实时控制方案。在战略层面，实时产生所有模式、所有班次的时刻表；在战术层面，采用驻站控制、车速调整、公交信号优先等方式，降低公交车运行轨迹与优化时刻表的差异。战略层面优化通过双层模型实现。控制方案实现了对多模式公交的鲁棒控制，可维持公交稳定运行，均衡公交载客量，快速消除异常事件的影响；（2）建立了考虑车内拥挤度的多模式公交组合优化设计模型，在模型中体现车内拥挤、找到座位的概率对乘客出行成本及乘客路径选择的影响。模型可更好地均衡不同模式、不同班次的客流分布，提高乘客找到座位的概率，改善乘客出行体验，降低公交系统总成本；（3）对影响公交运行的诸多因素进行了量化分析，鉴别出影响公交运行的关键因素及因素组合；（4）提出了基于APC数据分布的公交客流OD矩阵估计的新模型算法，有效提高了公交客流OD估计的精度；（5）提出了Wi-Fi探针及电子收费箱数据结合的公交载客量及客流OD估计的方法，为公交客流采集提供了新的思路。项目的研究成果可应用于公交系统的实际运营中，有利于提高公交运营的效率，同时提升乘客出行质量。