基于Q-learning和BP神经元网络的交叉口信号灯控制

解决单交叉口信号灯最优控制问题。提出了基于强化学习的信号灯控制系统结构,应用强化学习中Q学习,将信号灯最优控制问题转变成是否切换运行相位的决策问题,提出了采用BP神经元网络实现Q学习的信号灯控制系统。应用微观交通仿真软件PARAMICS进行仿真分析,结果表明该系统能够感知交通流变化,并能够自适应地调整信号灯切换策略,以达到最优的控制效果,该方法是可行的,与定时控制相比具有明显的优势。     

A comparison between PARAMICS and VISSIM in estimating automated field‐measured traffic conflicts at signalized intersections

The main objective of this study is to investigate the relationship between field‐measured conflicts and simulated conflicts estimated from microsimulation model (PARAMICS) using the surrogate safety assessment model. An urban signalized intersection was selected for analysis. Automated video‐based computer vision techniques were used to identify field conflicts. The applicability of a two‐step model calibration procedure applied to VISSIM in a recent study was investigated using PARAMICS. In the first calibration step, the PARAMICS model was calibrated to ensure that the simulation gives reasonable results of average delay times. The second calibration step used a genetic algorithm procedure to calibrate PARAMICS parameters to enhance the correlation between simulated and field‐measured conflicts. Finally, the results obtained from PARAMICS were compared with results obtained from VISSIM. The comparison included three aspects: (i) the car‐following model and safety‐related parameters; (ii) the correlation between simulated and field‐measured conflicts; and (iii) the conflict spatial distributions. The results show that the default simulation model parameters give poor correlation with the field‐measured data, and therefore, using simulation models without a proper calibration should be avoided. Overall, good correlation between field‐measured and simulated conflicts was obtained after calibration for both models, especially at higher time‐to‐collision (TTC) values. At TTC threshold of 1.5 s, PARAMICS overestimates the number of conflicts and VISSIM underestimates it. Both models overestimated the number of conflicts at TTC threshold of 3.00 s. There were major differences between field‐measured and simulated conflicts spatial distributions for both simulation models. This indicates that despite the good correlation obtained from the calibration process, both PARAMICS and VISSIM do not capture the actual conflict occurrence mechanism. Copyright © 2016 John Wiley & Sons, Ltd.     

基于多Agent的城市快速路协调控制研究

探讨了多Agent技术的特点及其应用,设计了基于多Agent的城市快速路控制系统的结构框架,包括区域协调Agent和控制Agent。并提出一种基于Q-学习的协调控制策略,实现Agent的学习功能。最后,通过微观交通仿真软件PARAMCIS对大望桥——四惠桥快速路交通区域建模,仿真结果表明结合Q-学习、模糊控制、仿真技术和信息诱导等方法的协调控制策略,可以在一定程度上改善整个快速路区域的交通状况。     

���п���·����ѵ������㷨��PARAMICS�����о�

本文基于对城市快速路入口匝道控制问题的分析，用微观交通仿真软件PARAMICS建立入口匝道控制仿真模型，对三种入口匝道控制算法（ALINEA，NEW-CONTROL和MIXED-CONTROL）进行仿真对比研究。结果表明三种算法都能一定程度上改善快速路的交通拥挤。同时给出三种算法的差异性描述，从交通流特征和算法特点分析产生差异的原因。最后指出MIXED-CONTROL在优化主线交通流的同时减少了入口匝道的车辆排队长度，其在高需求时的总体性能优于其它两种算法，是比较适合于实际应用的匝道感应控制算法。     

应用PARAMICS微观交通仿真对交通分流计划的比较评价

交通仿真日益被应用于ITS项目评价。本文提出一个基于PARAMICS微观交通仿真对ITS项目子系统和交通管理措施进行模拟评价的思路,并结合对交通拥挤管理措施之一的交通分流的实例评价进行论述和说明,在文末给出了比较和评价的结果。     

基于微观交通仿真的城市快速路限速控制方案评价

提出了基于PARAMICS仿真平台的限速控制方案的评价框架,应用PARAMICS模拟限速控制方案,并编写了模拟不同限速控制的程序插件,实例研究了发生事故的情况下,限速方案的仿真、结果分析和评价。     

An integrated modeling approach for facilitating emission estimations of alternative fueled vehicles

This paper develops an integrated model for reliable estimation of daily vehicle fuel savings and emissions using an integrated traffic emission modeling approach created by incorporating the US Environmental Protection Agency’s vehicle emission model, MOVES, and the PARAMICS microscopic traffic simulation package. A case study is conducted to validate the model using a well-calibrated road network in Greenville, South Carolina. For each transportation fuel considered, both emission and fuel consumption impacts are evaluated based on market shares.     

快速公交系统综合效果分析

近年来,国内很多城市都在致力于发展快速公交系统（BRT）,而正确认识BRT实施的综合效果是急需解决的关键问题。本文建立了一个通用的公共交通系统综合效果分析模型,采用综合赋权法标定参数,并以大连市华北路公交走廊为案例,对四种公共交通系统（普通公交、BRT、地铁、轻轨）进行数值计算分析。研究结果表明,BRT系统具有较好的经济效益和交通服务作用,可以在一定程度上替代轨道交通。本文建立的分析模型可以全面有效地评价BRT系统的综合效果,为城市公共交通管理者的决策提供理论依据与技术支撑。