Journal of Transportation Engineering, Part A: Systems

Both crash frequency analysis (CFA) and real-time crash prediction models (RTCPMs) divide a highway into small segments with a constant length [typically 0.161 km (0.10 mi)] for data aggregation. Many previous studies refer to this constant length as the segment length for data aggregation, but this paper adopts fragment size to avoid confusion with aggregation based on highway geometric features. Several studies have shown that segmentation length impacts the studies’ results and recommend not using a length smaller than 0.161 km (0.10 mi) or greater than 0.402 km (0.25 mi) to segment and aggregate traffic data for urban/suburban highways and freeways. Despite the significant impact of the segmentation length on traffic crash aggregation, no specific recommendation for selecting or determining the segmentation length for crash data aggregation exists. This research investigates the impact of segmentation length on traffic crash data aggregation. It establishes a methodology for determining a recommended fragment size (RFS) using hidden heterogeneity in traffic crash data. The study defines featured traffic crash rates using three major traffic crash characteristics: number of vehicles in crash, manner of collision, and crash severity. The analysis uses the Laplacian score with distance-based entropy measure and K-means to cluster highway segments based on the featured crash rates (FCRs) and total crash rates (TCRs) for fragment sizes ranging from 0.161 to 0.402 km (0.10 to 0.25 mi) with an increment of 0.016 km (0.01 mi). The clustering results are compared using their silhouette coefficients. The sample results shows that FCR-based clustering outperforms TCR-based clustering by providing important traffic crash groups within a highway and the RFS to segment and aggregate traffic crash data. The proposed method provides a data-driven comparison of different fragment sizes, revealing the pattern of traffic crashes and a standardized approach for RFS, which reduces the likelihood of fragment misclassification and benefits traffic studies depending on segmentation length.

Expansion of the rail transit network has increased public transportation accessibility, and the spatial and temporal distribution characteristics of ridership have become more complex. Herein, the characteristics of rail transit origin-destination (OD) ridership distribution were analyzed using a drift power law distribution model, and the robustness of ridership distribution in three periods (weekdays, weekends, and holidays) was investigated. A Louvain algorithm-based community detection method divided the rail transit network into different community structures. This division was generally consistent with the established rail transit line layout. However, central city communities were more dispersed, and the rail transit stations of different lines may have belonged to the same community, revealing the spatial correlation between different stations from the perspective of OD ridership. The outcomes of this division were stable for the different periods. This study contributes to the precise configuration of transportation facilities around rail transit stations and the formulation of rail transit operation management strategies.

Park-and-ride (P&R) facilities, when combined with the advantages of urban rail transit, have the potential to alleviate congestion problems and promote multimodal transportation. However, some facilities face operational and utilization issues in terms of efficiency and capacity, which may not fully meet users’ needs. Therefore, it is essential to conduct research that addresses challenges related to user parking and facility operation. A critical step in this direction is to study relevant indicators that comprehensively measure the operational efficiency and capacity of the facilities, and to find effective methods to tackle problems in P&R operation. This study examined and selected data indicators for P&R facilities around targeted stations, taking those around Nanjing urban rail transit stations as an example. Data envelopment analysis was employed to establish a P&R operational efficiency and capacity evaluation model. This model was enriched with a comprehensive evaluation of multiple factors that affect P&R, with these factors significantly influencing the recorded evaluation results. By identifying operational issues in P&R facilities around Nanjing urban rail transit stations, this study proposed optimization solutions to enhance their operational efficiency and capacity. The findings can serve as a reference for the coordinated development of P&R facilities and urban rail transit.

Rural communities are considered disadvantaged communities as they suffer from a lack of transport options. Thus, rural regions provide less accessibility for commuters to reach their destination as opposed to urban regions. However, the issues of transport disadvantage and shared use mobility in rural areas within the United States (US) have not been well investigated. Furthermore, transport disadvantage differs between communities and regions across the globe; thus, there is a need to study the behavioral choices of rural commuters within the US context. This study contributes by analyzing the behavioral choices of rural communities within the US through a case study site of Waynesburg, Pennsylvania, for adopting a shared use shuttle service. K-means clusters showed that trips from the survey data were a good representation of real trips from Ecolane. Furthermore, random parameter-based binary logit models were calibrated using data collected from students, faculty, and residents in Waynesburg, Greene County, to study the behavioral choices of commuters. The findings for the faculty and students group revealed that prior experience with shared services increases the likelihood of using a shared shuttle. An important personal characteristic of inconvenience showed a higher propensity toward using existing modes as opposed to a shared shuttle. Such commuters value personal vehicles as more convenient as they have childcare responsibilities and varying schedules for work that require them to move back and forth across locations, thus making a shared shuttle less attractive for them. The socioeconomic factors of age and gender show a higher propensity for using shared shuttles. The findings from this study could be helpful for agencies in improving rural mobility and considering such shared mobility services for rural communities.

Understanding charging vehicles, charging stations, and built environment concerning idle connection times significantly guided the management of charging infrastructure. However, the interplay between these factors had remained incompletely understood. This study addressed this gap by investigating public charging stations in Changshu, Suzhou, China, as a case study. The random forest regression and partial dependence plots were employed to explore the nonlinear relationships between idle connection times of vehicles at public fast-charging stations and the built environment, charging stations, and charging vehicles. The exploration encompassed two typical scenarios: workdays and weekends. The findings reveal the distinct influences of various factors in different scenarios. Notably, catering service points of interests in the proximity of charging stations, significantly impact the idle connection time on both workdays and weekends. Furthermore, government groups and residential areas have a notable influence on idle connection times during workdays. Shopping service and Leisure sport have a significant impact on idle connection time during the weekends. Variables such as the charging start time and charged energy also exhibit significant effects. Importantly, these influencing factors demonstrate heterogeneity and exhibit different threshold effects. This research can offer valuable insights to planning authorities and charging facility operators for formulating strategies to enhance charging infrastructure utilization.

The current arterial signal coordination models only target through traffic or critical traffic, resulting in fairness issues for traffic in other directions. Moreover, these models fail to achieve system optimality between intersections along the arterial. To address these problems, this paper establishes an arterial signal coordination optimization model based on all-direction pairs of adjacent intersections. The model takes the minimum weighted green wave band center offset of all-direction pairs at the intersections as the objective function. The optimization variables are the cycle length, offset, and phase sequence. The phase sequence structure adopts the ring-and-barrier structure, and the phase sequence optimization includes not only the main road but also the minor road intersecting with it. The model is an integer linear programming model, and it optimizes all-direction pairs of the upstream and downstream intersections. By analyzing the spatiotemporal diagram of the intersections, constraints are established for the phase sequence of the main road and intersecting roads. Through optimization of the offset and phase sequence, the coordination of flow direction pairs between upstream and downstream intersections can be optimized to improve the overall traffic fairness of the arterial.

Accurate detection of vehicles at nighttime is essential for transportation monitoring and management. However, annotating nighttime vehicle data is challenging, and vehicle features differ significantly between day and night, introducing difficulties in nighttime detection using pretrained models trained on daytime data. In this study, the nighttime vehicle detection performance is improved by employing a patchwise contrastive learning technique to enhance the representation of informative features for various traffic instances. An object detection network with reduced computational complexity and hyperparameters is utilized to conduct vehicle detection at night. Extensive experiments have been performed using images acquired from a section of Jingshi Road in Jinan, China. The impacts of learning rates and crop sizes are discussed. Three commonly adopted indicators, including mean average precision (mAP), precision, and recall, have been used to evaluate the training performance of the adopted FreeAnchor detector. Experimental results indicate that using a crop size of 320 and a learning rate of 2e-4, the developed generative adversarial network (GAN) achieves the best performance in image translation. Moreover, with a ratio of 60% real images to 40% fake images in model training, the FreeAnchor detector achieves the highest mAP of 96.6%. Visualized results for both image translation and nighttime vehicle detection demonstrate improved performance, underscoring the effectiveness of the proposed framework. This study paves the way for leveraging GAN-based networks to assist in vehicle detection under nighttime conditions.

Design guidelines for freeway ramp entrances are based on speed and acceleration data collected before 1950. This study investigated driver behavior over the entire freeway entrance area, including the ramp, the acceleration speed-change lane (SCL), and the freeway right lane (FRL). Video-based trajectory and speed profile data were collected using unmanned aerial vehicles (UAVs) and were used for qualitative and quantitative analysis. General trends of the relationships between driver behavior measures and geometric characteristics of entrance ramp terminals were investigated under different traffic and design conditions. Results showed that vehicles tended to merge onto the freeway at relatively low speeds such that the difference between their mean speed at merging and that of FRL vehicles was statistically significant. Results also confirmed that SCL drivers tended to start acceleration after they passed the middle of the ramp controlling curve. Regression models were developed for predicting driver-vehicle behavior on SCLs and on-ramp curves using traditional regression for each parameter separately and simultaneous modeling using structural equation modeling. An example application is presented to demonstrate the use of the developed models in reliability analysis of entrance ramps, which can be used to establish probabilistic road design guidelines.