Urban traffic volume detection is an essential part of trafficplanning in terms of urban planning in China. To improve the statisticsefficiency of road traffic volume, this thesis proposes a method for predictingmotor...Urban traffic volume detection is an essential part of trafficplanning in terms of urban planning in China. To improve the statisticsefficiency of road traffic volume, this thesis proposes a method for predictingmotor vehicle traffic volume on urban roads in small and medium-sizedcities during the traffic peak hour by using mobile signal technology. Themethod is verified through simulation experiments, and the limitations andthe improvement methods are discussed. This research can be divided intothree parts: Firstly, the traffic patterns of small and medium-sized cities areobtained through a questionnaire survey. A total of 19745 residents weresurveyed in Luohe, a medium-sized city in China and five travel modes oflocal people were obtained. Secondly, after the characteristics of residents’rest and working time are investigated, a method is proposed in this studyfor the distribution of urban residential and working places based on mobilephone signaling technology. Finally, methods for predicting traffic volume ofthese travel modes are proposed after the characteristics of these travel modesand methods for the distribution of urban residential and working placesare analyzed. Based on the actual traffic volume data observed at offlineintersections, the project team takes Luohe city as the research object and itverifies the accuracy of the prediction method by comparing the predictiondata. The prediction simulation results of traffic volume show that the averageerror rate of traffic volume is unstable. The error rate ranges from 10% to 30%.In this thesis, simulation experiments and field investigations are adopted toanalyze why these errors occur.展开更多
The Transport Authority of Clermont-Ferrand(France)has been facing the challenge of overcrowding on its tramway line for many years.It therefore considered an innovative approach to tackle the issue,by focusing on dem...The Transport Authority of Clermont-Ferrand(France)has been facing the challenge of overcrowding on its tramway line for many years.It therefore considered an innovative approach to tackle the issue,by focusing on demand management.The objective is to convince some organizations located along the line to change their working arrangements,so that their employees would avoid travelling during peak hours on the line.TTK,a German-French mobility planning and consulting company,was in charge of studying the opportunity and feasibility of this approach.For that purpose,a model was developed to assess how changes could improve the line overcrowding(or,how they could worsen the situation if not coordinated).This model was developed in-house,as traditional macro-simulation tools cannot be precise enough to understand the impact of 5 or 10 min changes on the line saturation.Model results identified several opportunities,and two organizations took first steps to implement change.Even with the sole participation of those two organizations,the SMTC saw a decrease of 5%demand at the peak times over a one-year period,a demand spread to off-peak times.This encouraging result led the SMTC to keep working with the organizations to develop this dynamic to other sites.展开更多
自主式交通系统(autonomous transportation system,ATS)是为应对主动式智慧交通发展趋势而提出的新一代交通系统。为科学合理地构建ATS功能架构,提出了一种面向多属性文本的优化密度峰值聚类算法(density peaks clustering,DPC)。该算...自主式交通系统(autonomous transportation system,ATS)是为应对主动式智慧交通发展趋势而提出的新一代交通系统。为科学合理地构建ATS功能架构,提出了一种面向多属性文本的优化密度峰值聚类算法(density peaks clustering,DPC)。该算法结合交通系统功能架构的基本特征,通过改进的词频-逆向文档频率算法与文本向量空间模型,将多属性文本转化成空间维度坐标。再利用高斯函数和决策值优化DPC算法进行聚类,并结合轮廓系数对聚类结果进行评价。为了检验算法的合理性,在ATS道路自动驾驶场景下,基于道路载运工具运行服务域、交通基础设施管理服务域和交通安全管理服务域的功能数据集进行了算例分析,依据聚类结果绘制功能架构图。架构图由自主感知-自主学习-自主决策-自主响应4层构成,验证了ATS应用场景中功能架构优化算法的可行性和合理性。算例结果表明:该算法的构建具有鲁棒性,算例轮廓系数整体均值为0.84,与原算法相比解决了聚类过程中聚类中心难以划定的问题;与原智能交通系统中的各架构设计相比,该功能架构更具有层次性和逻辑性。该优化算法能够促进新一代交通系统功能架构的构建,推动自主式交通系统理论体系的发展。展开更多
文摘Urban traffic volume detection is an essential part of trafficplanning in terms of urban planning in China. To improve the statisticsefficiency of road traffic volume, this thesis proposes a method for predictingmotor vehicle traffic volume on urban roads in small and medium-sizedcities during the traffic peak hour by using mobile signal technology. Themethod is verified through simulation experiments, and the limitations andthe improvement methods are discussed. This research can be divided intothree parts: Firstly, the traffic patterns of small and medium-sized cities areobtained through a questionnaire survey. A total of 19745 residents weresurveyed in Luohe, a medium-sized city in China and five travel modes oflocal people were obtained. Secondly, after the characteristics of residents’rest and working time are investigated, a method is proposed in this studyfor the distribution of urban residential and working places based on mobilephone signaling technology. Finally, methods for predicting traffic volume ofthese travel modes are proposed after the characteristics of these travel modesand methods for the distribution of urban residential and working placesare analyzed. Based on the actual traffic volume data observed at offlineintersections, the project team takes Luohe city as the research object and itverifies the accuracy of the prediction method by comparing the predictiondata. The prediction simulation results of traffic volume show that the averageerror rate of traffic volume is unstable. The error rate ranges from 10% to 30%.In this thesis, simulation experiments and field investigations are adopted toanalyze why these errors occur.
文摘The Transport Authority of Clermont-Ferrand(France)has been facing the challenge of overcrowding on its tramway line for many years.It therefore considered an innovative approach to tackle the issue,by focusing on demand management.The objective is to convince some organizations located along the line to change their working arrangements,so that their employees would avoid travelling during peak hours on the line.TTK,a German-French mobility planning and consulting company,was in charge of studying the opportunity and feasibility of this approach.For that purpose,a model was developed to assess how changes could improve the line overcrowding(or,how they could worsen the situation if not coordinated).This model was developed in-house,as traditional macro-simulation tools cannot be precise enough to understand the impact of 5 or 10 min changes on the line saturation.Model results identified several opportunities,and two organizations took first steps to implement change.Even with the sole participation of those two organizations,the SMTC saw a decrease of 5%demand at the peak times over a one-year period,a demand spread to off-peak times.This encouraging result led the SMTC to keep working with the organizations to develop this dynamic to other sites.
文摘自主式交通系统(autonomous transportation system,ATS)是为应对主动式智慧交通发展趋势而提出的新一代交通系统。为科学合理地构建ATS功能架构,提出了一种面向多属性文本的优化密度峰值聚类算法(density peaks clustering,DPC)。该算法结合交通系统功能架构的基本特征,通过改进的词频-逆向文档频率算法与文本向量空间模型,将多属性文本转化成空间维度坐标。再利用高斯函数和决策值优化DPC算法进行聚类,并结合轮廓系数对聚类结果进行评价。为了检验算法的合理性,在ATS道路自动驾驶场景下,基于道路载运工具运行服务域、交通基础设施管理服务域和交通安全管理服务域的功能数据集进行了算例分析,依据聚类结果绘制功能架构图。架构图由自主感知-自主学习-自主决策-自主响应4层构成,验证了ATS应用场景中功能架构优化算法的可行性和合理性。算例结果表明:该算法的构建具有鲁棒性,算例轮廓系数整体均值为0.84,与原算法相比解决了聚类过程中聚类中心难以划定的问题;与原智能交通系统中的各架构设计相比,该功能架构更具有层次性和逻辑性。该优化算法能够促进新一代交通系统功能架构的构建,推动自主式交通系统理论体系的发展。
