Peifeng Hu

Capacities of Unsignalized Intersections Under Mixed Vehicular and Nonmotorized Traffic Conditions

Unsignalized intersections consist of three types–-two-way stop-controlled, all-way stop-controlled, and uncontrolled intersections–- all with different priority relationships between traffic movements according to traffic laws. A conflict technique method was used to develop capacity models for the three types of unsignalized intersections under mixed traffic conditions involving vehicular, bicycle, and pedestrian movements. With field data collected from several unsignalized intersections, the model parameters were calibrated by a comparison analysis of traffic conditions in China and were modified on the basis of actual traffic conditions. The capacities obtained by the proposed models matched well with the observed capacities and the capacities calculated by conventional methods, both of which verified the effectiveness of the proposed models. The models proved to be valuable tools for determining capacities of vehicular movements at unsignalized intersections.

Field Evaluation of SCATS Control System in Las Vegas

This paper documents the results of a before-after study evaluating the effectiveness of Sydney Coordinated Adaptive Traffic System (SCATS), an adaptive traffic signal control system, implemented on Boulder Highway, an arterial in the Las Vegas urban area. The evaluation was based on comparisons of various performance measures before and after the SCATS system was implemented. Extensive travel time runs were conducted for all the major travel routes within the study network during the weekday and weekend peak periods. Travel times and number of stops were then compared between conventional time-of-day (TOD) coordinated control plans and SCATS control plans. Based on the data collected in this study, the general conclusion was that no significant improvement on arterial progression was achieved with SCATS under normal traffic conditions. However, this conclusion was made based on a major limitation of using video detection in the field, which may have significantly limited the performance of SCATS.

Study on Service-Level Classification of Platforms in Beijing Urban Rail Transit

The classification standards of levels of service (LOSs) for urban rail transit platforms are the foundation of platform design and evaluation. To study platform LOSs in urban rail transit, the passenger perception method and stated preference investigation concept were adopted, and a questionnaire was designed to investigate the factors influencing LOSs. The survey results showed that the congestion level of the platform was the most important factor influencing the LOS of the platform, followed by passenger order, air quality, information signs, and waiting time. On the basis of the results of the survey, the congestion level was used as the basis for the development of a LOS standard for platforms. A stated preference questionnaire was designed to investigate platform LOSs by using photos showing different congestion levels. According to the survey results, the membership curve of different LOSs was drawn, and the perceived LOS of the platform was classified into five levels on the basis of the congestion level. With linear interpolation, critical values of pedestrian density corresponding to different LOSs were calculated. Finally, the LOS standards developed for transit platforms were compared with the LOS criteria for pedestrian queuing areas offered by the Highway Capacity Manual, and possible reasons for the differences were analyzed. The proposed LOS standards for urban rail transit platforms can be used as a reference for transit design and evaluation.

Application of Combined Traffic Demand Forecasting for Comprehensive Transport Corridor

Traffic demand forecasting is the indispensable process of capacity and resource optimal allocation of comprehensive transport corridor. As such, it is derivate by social and economy activities. The steps of traffic demand forecast of comprehensive transport corridor in this paper is as follows; Step1: Using principal component analysis as a pretreatment to filter out the critical factors that affect traffic demand in corridor; Step2: Taking correlation coefficient method to forecast traffic demand in corridor; Step3: Forecasting based on BP Neural network; Step4: Adopting variance-covariance method to integrate above two forecasting results, then the result is the weighted average of the two methods. It is proved that not only vertical trend demand based on time sequence but also the horizontal demand based on social economic factors are taken into account in this model. Moreover, the deficiency of each method is overcomed and the advantages of these methods are integrated at the same time, which could improve the accuracy of forecasting and set up a solid basis for next capacity optimization.

Short-Term Traffic Flow Prediction Based on Time Series Analysis

During the past few years, extensive studies have been conducted on short-term traffic flow (STTF) prediction. Up to now, a large number of models have been proposed in this field. Among them, one kind of time series analysis, Auto Regressive Integrated Moving Average (ARIMA) with First Order Difference, is popular in STTF prediction. However, based on weekly similarities in traffic flow, a new Zero-Mean-Valued method is introduced in this paper. Then, an improved time series analysis approach is provided. The experiment was performed using inductive-loop data collected in five-minute intervals on Xianqian Street in Wuxi, China. A comparison was made between the results using First Order Difference and those using the approach proposed in this paper. In the end, the experiment results further proved that proposed method produced more consistent and more accurate forecasting results than the method using First Order Difference.