International Journal of Transportation Science and Technology最新文献

When there are multiple lanes to choose from downstream of a turning movement, drivers should choose the innermost lane so that drivers at other approaches of the intersection may make concurrent turning movements in the outermost lane(s). However, human drivers do not always choose the innermost lane, which could lead to crashes with other vehicles. Therefore, predicting human driver behaviors is vital in reducing crashes, as the need to share the roadways with automated vehicles (AVs) continues to grow. In this research, various machine learning models have been used to predict the left turn destination lane choice of human-driven vehicles (HDVs) at urban intersections based on several quantifiable parameters. A total of 174 subject vehicles were extracted and analyzed in Los Angeles, California, and Atlanta, Georgia, using HDV trajectory data from the Next Generation SIMulation (NGSIM) database. Five machine learning techniques, namely binary logistic regression, k nearest neighbors, support vector machines, random forest, and adaptive neuro-fuzzy inference system, were applied to the extracted data to predict the lane choice behavior of drivers. The k nearest neighbors model showed the most promising results for the evaluated data with a correct decision score of over 93% for the unseen test data. This model may be programmed into: (i) AVs, in conjunction with sensors, to predict if an HDV is about to turn into the incorrect destination lane; and (ii) microscopic traffic simulation tools so that modelers can identify potential conflicts when HDVs do not select the appropriate destination lane.

Adaptive Traffic Signal Control (ATSC) adjusts signal timings to real-time traffic measurements, increasing operational efficiency within a network. However, ATSC is both expensive to install and operate making it infeasible to deploy at all signalized intersections within a network. This study presents a bi-level optimization framework that applies heuristic methods to identify a limited set of locations for ATSC deployment within an urban network. At the upper-level, the Population Based Incremental Learning (PBIL) algorithm is employed to generate, evaluate, learn, and update different ATSC configurations. The lower-level uses the delay-based Max-Pressure algorithm to simulate the ATSC configuration within a microsimulation platform. The study proposes improvements to the PBIL algorithm by considering constraints on the maximum number of intersections for ATSC deployment and incorporates prior information about the intersection performance (i.e., informed search). Simulation results on the traffic network of State College, PA reveal that the proposed PBIL algorithm consistently outperforms baseline methods that select locations only based on queue-lengths or delays in terms of reducing overall network travel times. The study also reveals that intersections experiencing the highest delays or longest queues are not always the best candidates for ATSC. Moreover, applying ATSC at all intersections does not always provide the best performance; in fact, ATSC applied to some locations could increase travel times by contributing additional congestion downstream. Additionally, the modified PBIL algorithm with the informed search strategy is more efficient at identifying promising solutions suggesting it can be readily applied to more generalized optimization problems.

This work attempts to understand how a customized bus (CB) operator decides to open or close a CB line. We look into the changes in the operation status of CB lines (i.e. reopening and closure) from one of the largest CB operators in Shanghai, China, with a 22-month consecutive observation ranging from January 2019 to October 2020. As all CB services were totally suspended at the beginning of 2020 due to the COVID-19 travel restriction and then gradually recovered in March 2020, we utilize this study period as a naturalistic observation experiment to investigate the changes in the operation status of each CB line before and after the travel restriction. Using the operation status at each month as the binary alternatives, the mixed logit models and the tree-based models with explainable machine learning techniques are respectively adopted to explore the factors that influence the decision-making process. The findings from both types of models are in general consistent. The results show that the characteristics of each CB line including the ridership, the length of the line, the closeness to charging stations, and the overlap of CB lines significantly impact the decisions. In addition, the land-use types around the CB stops and the market competition from alternative travel modes also play a key role in making the decisions.

Since the adoption rate of e-grocery skyrocketed in the wake of the Covid-19 pandemic due to the influx of first-time e-grocery shoppers, grocery shopping behavior has been evolving and the travel effects of e-grocery are largely unknown. Thus, this study sought to examine the relationship between consumers’ grocery shopping behavior online and in-store, and the influencing factors (i.e., socio-demographic characteristics, household attributes, and personal attitudes). To achieve this, information relating to online and in-store grocery purchase frequencies, personal and household characteristics, and attitudes of more than 2,000 Florida residents were collected through an online survey. Using a bi-directional structural equation modeling (SEM) approach, our results show that online grocery shopping exhibited no significant effect on in-store grocery shopping frequency (i.e., neutrality), but in-store grocery shopping reduced the frequency of online grocery shopping (i.e., substitution). Also, a positive attitude toward some positive aspects of online shopping, preference for alternative travel modes, and tech savviness were associated with more frequent online grocery shopping, while cost consciousness and the joy of shopping encouraged more in-store shopping. Several socio-demographic and household attributes were also found to have direct and indirect effects mediated via attitudes on the shopping frequencies. Overall, this study provides insights into the demand and travel effects of e-grocery and highlights the need for retailers and transport planners to collaborate in order to mitigate the potential travel effects of e-grocery.

Connected and automated vehicles (CAVs) have the potential to revolutionise the transportation industry, with a plethora of research already revealing considerable gains in safety, travel time and mobility, as well as reduced congestion and pollution. As the number of CAVs on the road grows, rigorous testing for various market penetration rates (MPRs) of CAVS is essential to determine under what conditions the benefits can be realised. For the studies investigating the impact of CAVs on travel time reliability specifically, the MPRs in which the network most thrives have been inconsistent. The majority of the research is concerned with highway networks with only a few travel time reliability studies that focus on urban networks. In this simulation study, the impact of varying MPRs of CAVs on travel time reliability is evaluated in an urban network for different traffic demands. Travel time reliability metrics are assessed, including the standard deviation, buffer time index and misery index. The study demonstrated that from 0% to 100% MPR, the overall weighted average travel time decreased by 28%, and the standard deviation of the weighted average travel time declined by 35%, highlighting the significant increase in travel time reliability. Travel time improvements were visible from the MPR of 10%; however, the reliability metrics highlighted the greatest benefits occurred at higher MPRs. This study presents valuable results about the reliability that CAVs can bring to urban networks during the fleet transition to CAVs.

With the rapid expansion of urban areas, intercity highways have become crucial for daily transportation. Traffic administrators and planners increasingly rely on evaluating highway traffic volume. This paper aims to investigate the relationship between various factors and intercity traffic volume, with a specific focus on exploring the advancing and lagging effects of weather conditions on traffic volume in the districts of urban agglomerations. Using multiple data sources in the Guangdong-Hong Kong-Macao Greater Bay Area, including weather factors (i.e., rain, temperature, wind, and visibility), traffic factors (i.e., total traffic volume and travel time), and other factors (i.e., node degree, hub cities, and time of day), a mixed geographically weighted regression (MGWR) model is applied to examine the spatial heterogeneity of these factors. The results show that intercity traffic volume is influenced by weather, traffic, and other factors. Additionally, the advancing and lagging effects of different weather factors exhibit spatial heterogeneity across districts. Moreover, the weather lagging effect has a more significant impact than the advancing effect on intercity traffic volume. These findings provide valuable insights into the impact of weather on intercity travel volume and offer precise traffic guidance for intercity travelers.

The safety of aircraft landing on wet runways is of great importance in runway risk management. In order to ensure landing safety on wet runways, real-time risk warning is required. This paper proposes a method to assess aircraft landing risk in real-time based on finite element-virtual prototype-machine learning co-simulation. Firstly, a tire-water film-runway finite element model was constructed, a virtual prototype model was built based on the Airbus A320 model, and the results of the tire-water film-runway local finite element dynamic analysis were transferred to the system simulation of the virtual prototype for co-simulation. Secondly, considering the influence of wet state parameters on the runway, a database of aircraft anti-skid failure risk was constructed, and three machine learning models were trained to predict aircraft landing risk. The results show that the Support Vector Machine (SVM) model has better generalization capability and should be used to predict the risk level of aircraft landing. The efficacy of the comprehensive taxiing model was validated using an empirical formula for determining the aircraft's landing distance on a wet runway. When an aircraft lands on a runway with an average water film thickness of 8 mm, the braking time is approximately 1.6 times longer than on a dry runway, and the braking distance is roughly 5.3 times greater than on a dry runway. Finally, a risk assessment example was provided: the entire process from landing information input to risk level output for the aircraft model took only 80 ms, which could provide an efficient and real-time aircraft landing risk assessment.

Survey of road cracks in a timely, complete, and accurate way is pivotal to pavement maintenance planning. Motivated by the increasingly heavy task of identifying cracks, researchers have developed extensive crack segmentation models based on Deep learning (DL) methods with significantly different levels of accuracy, efficiency, and generalizing capacity. Although many of the models provide satisfying detection performance, why these models work still needs to be determined. The objective of this study is to survey recent advances in automated DL crack recognition and provide evidence for their underlying working mechanism. We first reviewed 54 DL crack recognition methods to summarize critical factors in these models. Then, we conducted a performance evaluation of fourteen famous semantic segmentation models using the quantitative metrics: F-1 score and mIoU. Then, the effective receptive field and class activation map of the included models are visualized to demonstrate the training results as qualitative evaluation. Based on the literature review and comparison results, larger kernel size, feature fusion, and attention module all contribute to the improvement of model performance. Striking a balance between increasing the effective receptive field and computational/memory efficiency is the key to designing DL crack segmentation models. Finally, some potential directions and suggestions for future development are provided, such as developing semi-supervised or unsupervised learning for the high cost of pixel-level labeling.

The ever-increasing number of accidents is detrimental to the sustainable objectives of work zones. Exploring the causes for work zone safety (WZS) issues, countermeasures, and effectiveness is significant to enhance the WZS. Therefore, this research aims to identify the factors influencing the WZS, identify the established countermeasures, examine the effectiveness of countermeasures, and develop a decision support framework to enhance WZS.

This study addresses the issue by utilizing the Scopus database and the VOSViewer data mining tool. A Bibliometric search followed by a Scientometric analysis was conducted to identify the highly researched areas. Qualitative content analysis was performed by exploiting the results, findings, and discussions from influential articles. A systematic coding process was carried out by employing Nvivo software to establish factors that influence the WZS and possible countermeasures.

The most influencing factors were driver speeding, inattention, non-compliance to traffic control measures, poor work zone (WZ) layout, lane closures, and adverse weather conditions. Highly discussed countermeasures include changeable/variable/dynamic message signs, advance traveler information systems (ATIS), channelizing devices, merge guidance and police enforcement presence. Further, a framework to support WZS decision-making is developed to assist industry practitioners in making informed decisions on WZS. The academic community can benefit from identifying the core literature related to urban WZS and knowledge gaps that need to address in the future.

To ensure the safety of urban rail transit operations and uncover the transmission dynamics of risk sources, a key risk chain mining method for urban rail transit operation is proposed. Firstly, the H-Apriori association rule algorithm is proposed for the characteristics of low frequency but high riskiness of high hazard degree risk sources in urban rail transit operation, which adds a new hazard degree evaluation index to the traditional Apriori algorithm and couples with support degree two-dimensionally to mine the strong association rules among risk sources. Secondly, we construct a weighted risk network with risk sources as network nodes and strong association rules as network edges, and propose a key risk chain mining method for urban rail transit operation based on path search theory to mine key risk chains from the weighted risk network. Finally, using the actual urban rail transit operation data of a city in China as an example, a total of 17 key risk chains are mined, and then 5 key risk sources and 8 key chain break locations are obtained by riskiness and frequency analysis of key risk chains, and control plans are proposed. The research outcomes introduce a novel approach to mining risk chains in urban rail transit operations, shedding light on the propagation mechanisms, triggering probabilities, and degrees of unsafety associated with risk sources. The results not only provide theoretical support but also offer methodological guidance for pinpointing locations of risk chain breaks and refining the control of risk sources.