Establishing comparison events/crashes is among the key challenges in safety analysis. This study proposes a spatial consideration for predicting scooter crashes using Utah's five years of crash data. It involves creating buffers ranging from 5 to 250 ft from the point of the scooter crash to obtain comparison crashes. The appropriate variables were selected based on the literature and engineering judgment. The Binary Logistic Regression was then applied to determine the appropriate buffer based on the consistency in the direction and magnitude of the impact of predictor variables. Results indicate that three variables, the junction type, lighting condition, and weather condition, are susceptible to changes in the direction of impact. Moreover, the study findings reveal that as the buffer distance increases, the magnitude of the impact of the variables decreases. Based on the results, a buffer of less than 50 ft is deemed appropriate for various analyses due to consistency in direction and the magnitude of impact. Further, the study findings show that intersections, dark-lighted conditions, summer season, and right-turning movements are more likely to be associated with scooter crashes. These findings can be crucial to transportation agencies and practitioners in improving the safety of scooter riders.
Persons with disabilities often struggle with accessible transportation. Research has shown that autonomous vehicles are a possible alternative mode of transportation for persons with disabilities. Information on the criteria for the pick-up and drop-off (PUDO) locations is critical for their success. The goals of this study were to identify the challenges for persons with disabilities with current PUDO zones, identify future needs for on-demand and autonomous vehicles, as well as users' comfort level with them. In-person and virtual interviews were conducted with 25 persons with mobility-related disabilities. The interviews contained both open-ended and multiple-choice questions with a focus on accessibility in transportation and autonomous vehicles. Participants reported positive views regarding autonomous vehicles, and their ability to alleviate the challenges with current modes of transportation. However, specific needs were identified for PUDO zones. The past experiences of participants, as well as the uncertainty that they may face in new scenarios influenced their responses and preferences for accessibility of PUDO zones as well as mechanisms for loading and unloading. This work identified several challenges with current PUDO zones and new information that can be used to inform decisions for the selection and criteria of PUDO zones.
To clarify the distribution characteristics of the external dimensions of two-axle trucks in the current market in China, and to provide a basis for revising relevant design standards and specifications for cargo vehicle parameters, statistical methods were used to analyze the correlation and distribution characteristics of the external dimension parameters of two-axle trucks. Based on the analysis results, combined with the types of vehicles used for highway tolls, representative models for two-axle trucks and parking garage design models were determined. The study results showed that there is a strong positive correlation between the length and wheel space of two-axle trucks, whereas the correlation between wheel space and body width or height is relatively weak. The external dimensions are ranked as follows: heavy trucks > medium trucks > light trucks > micro trucks, with micro trucks having a significant difference in body width compared to other types of goods vehicles. Based on the data, it is suggested that the representative model for Type I trucks be determined as a light truck with external dimensions of 6.0 m × 2.55 m × 3.6 m, and the representative model for Type II trucks is a two-axle heavy truck with external dimensions of 12 m × 2.55 m × 4 m. The representative vehicle types and their external dimensions for truck parking garage design are as follows: micro trucks − 6 m × 2.1 m × 2.7 m, Type I trucks − 7.5 m × 2.55 m × 3.5 m, Type II trucks − 12 m × 2.55 m × 4 m.
Modern surveillance systems increasingly adopt artificial intelligence (AI) for their automated reasoning capacities. While AI can save manual labor and improve efficiency, addressing the ethical concerns of such technologies is often overlooked. One of these AI application technologies is vehicle re-identification - the process of identifying vehicles through multiple cameras. If vehicle re-identification is going to be used on and with humans, we need to ensure the ethical and trusted operations of these systems. Creating reliable re-identification models relies on large volumes of training datasets. This paper identifies, for the first time, limitations in a commonly used training dataset that impacts the research in vehicle re-identification. The limitations include noises due to writing on images and, most importantly, visible faces of drivers or passengers. There is an issue if facial recognition is indirectly performed by these black box models as a by-product. To this end, an approach using an image-to-image translation model to generate less noisy training data that can guarantee the privacy and anonymity of people for vehicle re-identification is proposed.
Electric vehicles (EVs) promise significant advancements, including high energy efficiency and the facilitation of grid-stabilizing technologies such as vehicle-to-grid. However, their increased adoption introduces challenges such as elevated congestion, compromised safety, and grid instability. These challenges stem from differences in acceleration and deceleration patterns between EVs and internal combustion engine vehicles (ICEVs), mismatches between charging station demand and grid supply, and potential cyberattacks on the communications of EVs with charging stations and local grids. To address these issues, novel mathematical and machine-learning models have been developed. These models incorporate both simulated and real-world traffic flow data, charging station distribution and utilization data, and in-vehicle energy management and driver assistance data. The outcomes include optimally planned routes for EVs to destinations and charging stations, stabilized power distribution systems during peak hours, enhanced security in EV-station-grid communication, more energy-efficient storage systems, and reduced range anxiety for EV drivers. This paper systematically reviews the emerging impacts of EVs on evolving transportation systems, highlighting the latest developments in these areas and identifying potential directions for future research. By reviewing these specific challenges and solutions, this paper aims to contribute to the development of more efficient and sustainable electrified transportation systems.
Traffic state data, such as speed, density, volume, and travel time collected from ubiquitous roadway detectors require advanced network level analytics for forecasting and identifying significant traffic patterns. This paper leverages diverse traffic state datasets from the Caltrans Performance Measurement System (PeMS) hosted on the open benchmark and achieved promising performance compared to well-recognized spatial-temporal prediction models. Drawing inspiration from the success of hierarchical architectures in various Artificial Intelligence (AI) tasks, cell and hidden states were integrated from low-level to high-level Long Short-Term Memory (LSTM) networks with the attention pooling mechanism, similar to human perception systems. The developed hierarchical structure is designed to account for dependencies across different time scales, capturing the spatial-temporal correlations of network-level traffic states, and enabling the prediction of traffic states for all corridors rather than a single link or route. The efficiency of the designed hierarchical LSTM is analyzed by ablation study, demonstrating that the attention-pooling mechanism in both cell and hidden states not only provides higher prediction accuracy but also effectively forecasts unusual congestion patterns. Data and code are made publicly available to support reproducible scientific research.
This paper conducts a thorough review of airline disruption management between 2010 and 2024. Unlike previous review papers, the present paper analyses the research on airline disruption management in three ways. One is to perform a statistical analysis of these papers based on the journal distribution, number of papers by year, and types of recovery resources. The second is to categorize integrated recovery methods based on the degree of integration of the resources during the recovery process: the aircraft and crew, the aircraft and passengers, and all three resources. The last way is to study the research findings based on statistical analysis and perform future research direction identification in the areas of problems, models, and solution approaches. Further, with the increasing complexity of actual demands, integrated flight disruption recovery considering multiple factors such as aircraft, crew, and passengers has become a research hotspot in recent years. For further research, we can delve deeper into issues from both practical circumstances and theoretical extensions. At the model level, more detailed characterizations are needed, along with more efficient solution methods to accommodate increasingly complex problems.
Timely and accurate traffic flow prediction is the core of an intelligent transportation system. Canonical long short-term memory (LSTM) networks are guided by the mean square error (MSE) criterion, so it can handle Gaussian noise in traffic flow effectively. The MSE criterion is a global measure of the total error between the predictions and the ground truth. When the errors between the predictions and the ground truth are independent and identically Gaussian distributed, the MSE-guided LSTM networks work well. However, traffic flow is often impacted by non-Gaussian noise, and can no longer maintain an identical Gaussian distribution. Then, a
