Dynamic graph convolutional network for long-term traffic flow prediction with reinforcement learning

We explore the impact of bicycle-sharing infrastructure on urban transportation. Accounting for selection bias in a matching framework, we estimate a causal effect of the Capital Bikeshare on traffic congestion in the metropolitan Washington, DC, area. We exploit a unique traffic dataset that is finely defined on a spatial and temporal scale. Our approach examines within-city commuting decisions as opposed to traffic patterns on major thruways. Empirical results suggest that the availability of a bikeshare reduces traffic congestion by 2 to 3% within a neighborhood. We also identify geographic spillovers that may counteract benefits from reductions in local pollution.

https://papers.ssrn.com/sol3/Delivery.cfm/SSRN_ID2649978_code1108470.pdf?abstractid=2649978&mirid=5&type=2

Bicycle Infrastructure and Traffic Congestion: Evidence from DC's Capital Bikeshare

https://scholarship.richmond.edu/cgi/viewcontent.cgi?article=1059&context=economics-faculty-publications

Quality of location-based crowdsourced speed data on surface streets: A case study of Waze and Bluetooth speed data in Sevierville, TN

To improve traffic operations and management under inclement weather, this work quantificationally investigated the influence of rainfall on the traffic flow characteristics of urban expressways in Beijing. Four basic traffic flow models (Greenshields model, Greenberg model, Underwood model, and Drake function model) were compared to determine the most suitable model for the detected data, which were collected by microwave detectors in the Beijing 2nd and 4thring expressway under normal or rainfall condition. Results showed that the Greenshields model was the best-fitted model. Under rainfall conditions, varying degrees of reductions were observed in the road capacity, free-flow speed, and speed at capacity. The findings also showed that the reductions of road capacity were 7.26%, 10.87%, and 17.09%, and the reductions of free-flow speed were 3.07%, 5.29%, and 6.64% under light rain, medium rain, and heavy rain, respectively. The reductions of speed at capacity revealed a similar tendency as that of the free-flow speed. Simultaneously, a new model that incorporated rainfall intensity is proposed on the basis of the Greenshields model and can achieve good performance. Overall, the results are beneficial for traffic managers in improving their decision
 s and predictions under rainfall weather.

Impact Analysis of Rainfall on Traffic Flow Characteristics in Beijing

Travel time data are a key input to applications of intelligent transportation systems. Advancement in vehicle tracking and reidentification technologies and proliferation of location-aware and connected devices have made networkwide travel time data available to transportation management agencies. The trend started with data collection on freeways and has been quickly extended to arterials. Although the freeway travel time data have been validated extensively in recent years, the quality of arterial travel time data is not well known. This paper presents a comprehensive validation scheme for arterial travel time data based on GPS probe and Bluetooth data as two independent sources. Since travel time on arterials is subject to a higher degree of variation than that on freeways, mainly because of the presence of signals, a new validation methodology based on the coefficient of variation is introduced. Moreover, a context-dependent travel time fusion framework is developed to improve the reliability of trav...

Arterial Travel Time Validation And Augmentation With Two Independent Data Sources

Arterial Travel Time Validation and Augmentation with Two Independent Data Sources

Optimally deploying sensors into a complex highway network can significantly enhance the socio-economic benefits through accurately providing system users and operators with the latest traffic state information. This paper presents a linear optimization framework for designing a sensor network for travel time surveillance. The presented framework consists of a traditional static network model and a novel dynamic network model as well as an integer cut-based speed-up solution approach. The aims of the study are twofold. First, the proposed optimization model maximizes the network-level travel time surveillance benefits by taking into account the values of real-time measurements induced from both temporal and spatial traffic state patterns. Second, the mechanism of sensor relocation operations is designed and incorporated into the proposed model. This mechanism enables one to evaluate whether appropriately relocating sensors can optimize the utilization of the surveillance resources and further enhance the surveillance benefits over a given time horizon. We applied the proposed optimization framework to a mixed arterial-freeway commuting network located in Washington. D.C.-Baltimore metropolitan area. The study empirically revealed and illustrated two interesting findings: (1) relocation operations do bring additional surveillance benefits on travel time prediction, and the marginal benefit is highly correlated with the sensor fleet size; (2) spatial measurement is not a trivial part in travel time prediction, and one should fully recognize the effectiveness of temporal and spatial measurements when designing a traffic sensor network for real-time travel time surveillance.

Is dynamic traffic sensor network profitable for network-level real-time information prediction?

The purpose of this paper is to quantify the impact of rainfall on freeway traffic flow using toll station data and precipitation data. Five rainfall categories were classified based on precipitation intensity. Hourly traffic volume and travel times were retrieved and analysed in three spatial scales: at toll stations, on freeway sections between two toll stations, and on freeway corridors composed of multiple sections. Travel times for passenger cars in each rainfall category were studied and the impact of rainfall on traffic volume and travel time was quantified. The non-parametric Kruskal-Wallis test was performed to check the significance of difference among various categories. Calculated traffic volume reduction coefficients and travel time incremental coefficients for each rainfall category were compared with existing research results, which showed obvious differences. Potential causes are discussed to explain the discrepancy in coefficient values.

Study of the impact of rainfall on freeway traffic flow in Southeast China

In many emergency management operations, an efficient evacuation strategy is of great importance because if it is successful, it has the ability to significantly reduce the loss of property and human life This paper develops a routing and scheduling optimization framework for large-scale vehicular evacuation To guarantee high optimization efficiency, we consider the routing and scheduling optimization as a two- stage problem instead of optimizing them as a whole (ie using time-space network) In the first phase, a multiple-objective mixed integer programming (MIP) model, with the objectives of minimizing the total in-network time and network clearance time is proposed to find an optimal routing plan In the second phase, a simulation-based scheduling Heuristic is proposed to dynamically generate the time-dependent departure rates A real-world evacuation scenario in Eastern Shore of Maryland is studied by using the proposed optimization model The calculation results indicate a good optimization capability and flexibility of the proposed model Keywords: Vehicular Evacuation, Routing, Scheduling, Optimization, Heuristic Copyright c 2016 SERSC Language: en

Xuechi Zhang

Papers

Arterial Travel Time Validation And Augmentation With Two Independent Data Sources

Arterial Travel Time Validation and Augmentation with Two Independent Data Sources

Is dynamic traffic sensor network profitable for network-level real-time information prediction?

Study of the impact of rainfall on freeway traffic flow in Southeast China

A two-stage optimization framework for optimizing large-scale vehicular evacuation routing and scheduling operation with uninterrupted traffic flow