Search

Highlights • Tweets are mapped into numerical feature vectors using word-embedding models. • Tweets are classified into non-traffic, traffic incident, and traffic information. • Classification task is performed using convolutional and recurrent neural networks. • 51,100 tweets are collected, labeled, and publicly released for future research. • Models' superiority is demonstrated through several evaluation steps. Abstract In recent years, several studies have harnessed Twitter data for detecting traffic incidents and monitoring traffic conditions. Researchers have utilized the bag-of-words representation for converting tweets into numerical feature vectors. However, the bag-of-words not only ignores the order of tweet's words but suffers from the curse of dimensionality and sparsity. A common approach in literature for dimensionality reduction is to build the bag-of-words on the top of pre-defined traffic keywords. The immediate criticisms to such a strategy are that the pre-defined set of keywords may not include all traffic keywords and the tweet language is subjected to change over time. To address these shortcomings, we utilize the power of deep-learning architectures for both representing tweets in numerical vectors and classifying them into three categories: 1) non-traffic, 2) traffic incident, and 3) traffic information and condition. First, we map tweets into low-dimensional vector space through word-embedding tools, which are also capable of measuring the semantic relationship between words. Supervised deep-learning algorithms including convolutional neural network (CNN) and recurrent neural network (RNN) are then deployed on the top of word-embedding models for detecting traffic events. For training and testing our proposed model, a large volume of traffic tweets is collected through Twitter API endpoints and labeled through an efficient strategy. Experimental results on our labeled dataset show that the proposed approach achieves clear improvements over state-of-the-art methods. [ABSTRACT FROM AUTHOR]

A critical deficiency in any one or a combination of three transportation system characteristics: the driver, roadway, or vehicle can contribute to an elevated crash risk for the motoring public. Traffic signs often convey critical information to drivers. However, traffic signs are only effective when clearly visible and legible. Traffic sign vandalism that is exclusively the results of humans causes both sign legibility and visibility to deteriorate. Transportation agencies spend a significant amount of money to repair or replace vandalized signs. This study was conducted to identify which traffic signs are more vulnerable to vandalism. To do this, a mobile-based vehicle collected data of over 97000 traffic signs managed by the Utah Department of Transportation (UDoT), US. The vandalized signs were identified by a trained operator through inspection of daytime digital images taken of each individual sign. Location data obtained from online sources combined with the traffic sign data were imported into ArcGIS to acquire localized conditions for each individual sign. According to the chi-square test results, the association between vandalism and traffic sign attributes and localized conditions, including background color, size, mount height, exposure, land cover, and road type was evident. After employing the random forests model, the most important factors in making signs vulnerable to vandalism were identified. [ABSTRACT FROM AUTHOR]

Identifying the distribution of users’ transportation modes is an essential part of travel demand analysis and transportation planning. With the advent of ubiquitous GPS-enabled devices (e.g., a smartphone), a cost-effective approach for inferring commuters’ mobility mode(s) is to leverage their GPS trajectories. A majority of studies have proposed mode inference models based on hand-crafted features and traditional machine learning algorithms. However, manual features engender some major drawbacks including vulnerability to traffic and environmental conditions as well as possessing human’s bias in creating efficient features. One way to overcome these issues is by utilizing Convolutional Neural Network (CNN) schemes that are capable of automatically driving high-level features from the raw input. Accordingly, in this paper, we take advantage of CNN architectures so as to predict travel modes based on only raw GPS trajectories, where the modes are labeled as walk, bike, bus, driving, and train. Our key contribution is designing the layout of the CNN’s input layer in such a way that not only is adaptable with the CNN schemes but represents fundamental motion characteristics of a moving object including speed, acceleration, jerk, and bearing rate. Furthermore, we ameliorate the quality of GPS logs through several data preprocessing steps. Using the clean input layer, a variety of CNN configurations are evaluated to achieve the best CNN architecture. The highest accuracy of 84.8% has been achieved through the ensemble of the best CNN configuration. In this research, we contrast our methodology with traditional machine learning algorithms as well as the seminal and most related studies to demonstrate the superiority of our framework. [ABSTRACT FROM AUTHOR]

This study presents a methodology in a disaggregated manner to assess the spatial distributional effects of implemented toll road networks. Existing zone-based research focuses only on revealing the existence of regressivity. This approach provides limited information in understanding the spatial distribution of travelers from low-income groups known to have suffered welfare loss due to toll roads. For road pricing to be a win-win strategy for all income groups, the revenue from toll roads should be used to improve the mobility of low-income groups. Therefore, it is desirable to have more detailed location information and regressivity effects for each income group due to toll roads. This study proposes a new method to better represent the spatial distributional effects by utilizing disaggregated and recently available data (e.g., smartphone navigation data), enabling analysts to easily estimate the welfare loss of low-income groups due to toll roads. Applying the proposed method in Chungcheongnam-do, Korea, reveals that installing toll roads would be regressive in terms of income. However, the analysis results provide sufficient information to interpret spatial distributional effects', current status, and causes. This method improves the existing zone-based analysis, which cannot distinguish who has greater welfare loss among high- and low-income groups in a particular zone. This information would be extremely useful in establishing mitigation policies for low-income groups affected by installing toll roads in the future. [ABSTRACT FROM AUTHOR]

Introduction : Animal–vehicle collisions (AVCs) can result in serious injury and death to drivers, animals' death, and significant economic costs. However, the cost effectiveness of the majority of AVC mitigation measures is a significant issue. Method : A mobile-based data collection effort was deployed to measure signs under the Utah Department of Transportation's (UDOT) jurisdiction. The crash data were obtained from the UDOT risk management database. ArcGIS was employed to link these two data sets and extract animal-related crashes and signs. An algorithm was developed to process the data and identify AVCs that occurred within sign recognition distance. Kernel density estimation (KDE) technique was applied to identify potential crash hotspots. Results: Only 2% of AVCs occurred within the recognition distance of animal crossing signs. Almost 58% of animal-related crashes took place on the Interstate and U.S. highways, wherein only 30% of animal crossing signs were installed. State routes with a higher average number of signs experienced a lower number of AVCs per mile. The differences between AVCs that occurred within versus outside of sign recognition distance were not statistically significant regarding crash severity, time of crash, weather condition, driver age, vehicle speed, and type of animal. It is more likely that drivers become accustomed to deer crossing signs than cow signs. Conclusions : Based on the historical crash data and landscape structure, with attention given to the low cost safety improvement methods, a combination of different types of AVC mitigation measures can be developed to reduce the number of animal-related crashes. After an in-depth analysis of AVC data, warning traffic signs, coupled with other low cost mitigation countermeasures can be successfully placed in areas with higher priority or in critical areas. Practical applications : The findings of this study assist transportation agencies in developing more efficient mitigation measures against AVCs. [ABSTRACT FROM AUTHOR]

In this paper, a methodology for accelerating the double-projection method designed for the asymmetric traffic equilibrium assignment problem is proposed. One of the factors that determine the efficiency of the projection method is a properly selected step size. This study customized the step size for each origin-destination (O-D) pair by formulating the assignment problem on the space of path flows. In addition, even if the transportation network is decomposed into each O-D pair, the step size based on the entire path flows is also available. The study developed a step-size selection strategy with two different available step sizes obtained from different spaces of path flows vectors. During the projections of path flows for each O-D pair, the strategy compares the step size customized for that O-D pair and the one from the entire path flow vectors and selected a bigger step size. Several numerical tests have been conducted on various sized networks and convergence criteria. The methodology considerably improves the convergence speed of the solution algorithm. Further analysis on the numerical tests delivered the relevant information on the synergistic effects of the two strategies during the convergence process: the customized step size and the step-size selection strategy. [ABSTRACT FROM AUTHOR]

In this paper, the crowding effect in a transit vehicle is modeled in a time-expanded network that considers the daily variation in passenger flows. The study models the daily variation of in-vehicle crowding in a real large-scale transit system. A transit assignment for this real network is modeled and implemented by constructing a crowding cost function that follows the valuation of crowding and by using the reliable shortest path finding method. The direct application of the crowding model to a real network for the Utah Transit Authority indicates that crowd modeling with multi-user classes could influence public transportation system planning and affect the revenues of transit agencies. Moreover, the addition of the disutility factor, crowding, does not always appear to cause an increase in disutility for transit users. [ABSTRACT FROM PUBLISHER]

To ensure traffic sign visibility for the information transfer between traffic signs and drivers, transportation agencies are required to achieve and maintain traffic signs to at least minimum retroreflectivity levels. To do this, the Manual on Uniform Traffic Control Devices (MUTCD) outlined two assessment methods – visual nighttime inspection and retroreflectivity measurement – and three management methods – expected life, blanket replacement, and control signs. This study aimed to examine whether or not visual condition derived from daytime digital imaging can be used to achieve sign retroreflectivity compliance. To accomplish this goal, we used two data sets. A sample data of over 1700 signs was manually collected in the field, and the background retroreflectivity of each sign was measured using a handheld retroreflectometer. In addition, sign data of over 97,000 traffic signs was digitally collected by driving an equipped vehicle. Sign visual condition ratings (good, fair, and poor) data was obtained from inspection of daytime digital images taken of each individual sign. By focusing on the overlaps in the two data sets including almost 1500 signs, we obtained a range of retroreflectivity measurements for each sign condition. Also, we employed the basic definition of conditional probability to assess the probability of traffic sign retroreflective pass or failure, depending on whether the sign was recorded in good, fair, or poor conditions. At its completion, the study found that though daytime digital imaging cannot be considered a reliable alternative for sign retroreflectivity measurements, nighttime digital imaging may have potential for use, and both allow transportation agencies to know which signs should be investigated manually. [ABSTRACT FROM AUTHOR]

Previous research has found that distracted driving measurably increases driver response times to unanticipated roadway hazards. These instances are of particular consequence as they tend to be highly correlated with vehicle crashes resulting in property damage and/or injury. However, comparatively little attention has been allocated to quantifying the negative impacts of distracted driving on driver reaction to anticipated stimuli. This study empirically determined the impact of distracted driving on queue discharge rates at signalized intersections with protected left-turn phases for dual left-turn lanes. Observational studies were conducted at 11 intersection approaches at six signalized intersections in three states resulting in the observation of 844 distracted and 3,726 undistracted left-turning drivers. A statistical model was developed for the start-up lost time resulting from the average headways of the samples in each state. The difference in start-up lost times between queues comprised entirely of distracted or undistracted drivers ranged from 3.36 to 4.06 s in Kansas, 2.97 to 4.41 s in Oregon, and 2.25 to 5.14 s in Utah. [ABSTRACT FROM AUTHOR]

The novel coronavirus (COVID-19) pandemic created an environment where nearly all aspects of mobility changed to ensure the health and safety of the public. The Centers for Disease Control and Prevention (CDC) recommended that people quarantine for 14 days if they were potentially exposed to the virus, stay at least six feet apart from others, and stay at home as much as possible. Delivery via third-party restaurant app, grocery, and package delivery quickly became an essential service. This study assesses customer's changes in use and perceived quality of delivery services in Southwest Virginia, via an online stated-preference survey (n = 423). The responses were analyzed using ordered logit and generalized ordered logit models to identify which population segments had changing delivery behavior and perceptions due to the pandemic. Findings include that before the pandemic, only households with an income greater than $100,000 had a significantly higher demand for package delivery services than those making less than $25,000. During the pandemic, all income brackets had a significantly higher demand for package delivery "weekly" than households with less than a $25,000 income, with a 19.50%, 22.54%, and 45.59% greater chance of use for income levels $25,000 to $50,000, $50,000 to $100,000, and over $100,000, respectively. This trend highlights that package delivery became necessary during the pandemic. Respondents who lived within town limits were statistically significantly more likely to use third-party restaurant delivery apps at least once a week before (3.10%), during (9.20%), and after (4.50%) the pandemic compared to those outside town limits. The results also found people who lived within town limits were 7.77% more likely to be satisfied with delivery services in general than those who lived outside town limits. The findings from this paper identify expanding delivery equity gaps within the population and provide recommendations for policymakers and delivery agencies. Some limitations include that low sample size did not allow for fully segmented models and meant that the study should be considered exploratory in nature. • A stated preference survey about delivery preferences and use is conducted. • The study evaluates how the COVID-19 pandemic changed delivery habits. • There is a predicted rise in food, grocery, and package delivery use post-pandemic. • The study reveals an inequity in delivery service access. [ABSTRACT FROM AUTHOR]

This research was conducted to determine how best to incorporate portable rumble strips into traffic control plans for short-term work zones. One objective of the research was to provide guidance on how existing traffic control plans could be amended to include provisions for the evaluated devices. This research focused on rumble strips that were considered beneficial for any type of work zone where the duration is one day or less. The tests on the portable rumble strips were conducted on a closed roadway surrounding the Kansas Speedway in Kansas City, Kansas, and in a closed park-and-ride parking facility in Lawrence, Kansas. The rumble strips tested in this study were two different types of reusable temporary rumble strips made out of steel with a rubber bottom and four generations of plastic rumble strips. From this study the best solution for most short-term work zones would be the latest generation of plastic rumble strips tested. The earlier generations did not perform as well as the fourth generation especially at 60 mph (96.6 km/hr). The steel rumble strips also hold some promise; however, the structural integrity of the steel rumble strips is an issue that needs to be addressed. [ABSTRACT FROM AUTHOR]

• This study examines self-reported responses of 4,706 drivers to a fabricated message. • Female and young respondents highly expect to detour, change speed, or be distracted. • Familiarity with DMS increases the chance of detouring or changing speed. • Respondents expect to be distracted cognitively, visually, and manually. • Tech friendly respondents expect to be distracted more by taking pictures, texting, and calling. This article seeks to understand the potential response of drivers when they encounter compromised Dynamic Message Signs (DMS). The findings are built on the self-reported response of 4,706 participants in a Stated Preference (SP) survey conducted between November 2018 and December 2018 in the United States. The findings show the response of drivers to the "Downtown Under Terrorist Attack" message falls into route divergence, speed change, and distraction, and the likelihood of route divergence, distraction, and a slowdown is significantly more than either stopping or speeding up. The possibility of a response, however, varies depending on socioeconomic characteristics, attitudinal characteristics, and driving behavior. It is highlighted (1) female and young drivers are more probable to detour, to change speed, or to be distracted, (2) drivers who are familiar with DMS, encounter it frequently, or pay attention to its content, have a higher chance of detouring or changing speed, while they are less likely to be distracted, and (3) technology-friendly drivers are likely to detour or slow down. From the distraction model, it is further inferred that drivers are distracted cognitively, visually, and manually. The findings have implications for researchers and federal, state, and local agencies who are aware of the consequences of cybersecurity threats for the operation and profitability of the transport network. They, for example, assist transport planners in prioritizing equipment security efforts and resource allocation to the areas of greatest risk, and help to prepare contingency plans based on drivers' behavioral response. [ABSTRACT FROM AUTHOR]

This study explores speed choice behavior of travelers under realistic and fabricated Dynamic Message Signs (DMS) content. Using web-based survey information of 4,302 participants collected by Amazon Mechanical Turk in the United States, we develop a set of multivariate latent-based ordered probit models participants. Results show female, African-Americans, drivers with a disability, elderly, and drivers who trust DMS are likely to comply with the fabricated messages. Drivers who comply with traffic regulations, have a good driving record, and live in rural areas, as well as female drivers are likely to slow down under fabricated messages. We highlight that calling or texting, taking picture, and tuning the radio are distracting activities leading drivers to slow down or stop under fictitious scenarios. [ABSTRACT FROM AUTHOR]

Identification of travelers’ transportation modes is a fundamental step for various problems that arise in the domain of transportation such as travel demand analysis, transport planning, and traffic management. In this paper, we aim to identify travelers’ transportation modes purely based on their GPS trajectories. First, a segmentation process is developed to partition a user's trip into GPS segments with only one transportation mode. A majority of studies have proposed mode inference models based on hand-crafted features, which might be vulnerable to traffic and environmental conditions. Furthermore, the classification task in almost all models have been performed in a supervised fashion while a large amount of unlabeled GPS trajectories has remained unused. Accordingly, we propose a deep SEmi-Supervised Convolutional Autoencoder (SECA) architecture that can not only automatically extract relevant features from GPS segments but also exploit useful information in unlabeled data. The SECA integrates a convolutional-deconvolutional autoencoder and a convolutional neural network into a unified framework to concurrently perform supervised and unsupervised learning. The two components are simultaneously trained using both labeled and unlabeled GPS segments, which have already been converted into an efficient representation for the convolutional operation. An optimum schedule for varying the balancing parameters between reconstruction and classification errors are also implemented. The performance of the proposed SECA model, trip segmentation, the method for converting a raw trajectory into a new representation, the hyperparameter schedule, and the model configuration are evaluated by comparing to several baselines and alternatives for various amounts of labeled and unlabeled data. Our experimental results demonstrate the superiority of the proposed model over the state-of-the-art semi-supervised and supervised methods with respect to metrics such as accuracy and F-measure. [ABSTRACT FROM AUTHOR]

This study explores the cycling usage and frequency determinants in college campuses located in the Baltimore Metropolitan Area. The study discerns the attitudes of individuals toward the proposed infrastructure and environmental improvements with the goal of promoting biking to campus. We develop a structural equation model (SEM) using the travel information of 780 individuals, which was collected between December 2014 and June 2015. The results indicate risk factors have a higher explanatory value on bike-to-campus frequency than campus infrastructure and program. We further examine how and to what extent mixed populations on college campuses respond to latent factors. The findings pinpoint that males are less concerned about the risk-related indicators such as theft and road and environment-related obstacles such as poor road conditions. However, females have a positive attitude toward campus-related improvements such as pro-bike programs. Overall, students show a negative attitude toward the road and environmentally-related obstacles compared to staff and faculty. Minority groups, specifically African American and Asian, show a positive attitude toward campus-related improvements, unlike white participants. The findings can assist planners and advocates in implementing effective policy measures to increase bike-to-campus frequency. • Minorities have a positive attitude toward campus-related improvements. • Students have a negative attitude toward road and environmentally-related obstacles, compared to staff and faculty • Females are more concerned about risk-related indicators and environment related obstacles. • Females have a positive attitude toward campus-related improvements such as pro-bike programs. • Strong/fearless and enthused/confident cyclists were less concerned about road and environmentally-related indicators. [ABSTRACT FROM AUTHOR]

Many cities around the world suffer from air pollution but Tehran is among the worst. Tehran suffers severely from congestion and air pollution, which is caused primarily by private vehicles. To mitigate these problems, Transportation Demand Management (TDM) policies should be instituted across the city to reduce private car usage. However, the acceptability of these policies is a major concern for policymakers. The goal of this research was to gauge individual opinions with regards to the increasing petrol price policy, followed by a comparison of these opinions with other TDM pricing policies, including cordon pricing or parking pricing. To do this, we conducted a survey of 366 car commuters in Tehran. Throughout the analysis of collected data, it was observed that increasing petrol price was the least acceptable of the above policies. In order to identify the effects of trip-related and socio-economic characteristics of car commuters on their acceptance of increasing petrol price, we developed an ordered logit model and identified the most significant factors affecting commuters. Next, we suggested policies that could lead to better acceptance of the increasing petrol price policy. This examination of the acceptability of TDM pricing policies helps policymakers to understand how pricing policies will have a greater chance of successful implantation and acceptance by the general public, thus allowing them to make informed policy implementation decisions. [ABSTRACT FROM AUTHOR]

• A CNN approach for classifying vehicles based on GPS data is proposed. • A novel representation of a GPS trajectory suitable for deep learning is proposed. • The approach outperforms traditional machine learning methods. • The approach increases usability of GPS trajectory data. Transportation agencies are starting to leverage increasingly-available GPS trajectory data to support their analyses and decision making. While this type of mobility data adds significant value to various analyses, one challenge that persists is lack of information about the types of vehicles that performed the recorded trips, which clearly limits the value of trajectory data in transportation system analysis. To overcome this limitation of trajectory data, a deep C onvolutional N eural N etwork for V ehicle C lassification (CNN-VC) is proposed to identify the vehicle's class from its trajectory. This paper proposes a novel representation of GPS trajectories, which is not only compatible with deep learning models, but also captures both vehicle-motion characteristics and roadway features. To this end, an open source navigation system is also exploited to obtain more accurate information on travel time and distance between GPS coordinates. Before delving into training the CNN-VC model, an efficient programmatic strategy is also designed to label large-scale GPS trajectories by means of vehicle information obtained through Virtual Weigh Station records. Our experimental results reveal that the proposed CNN-VC model consistently outperforms both classical machine learning algorithms and other deep learning baseline methods. From a practical perspective, the CNN-VC model allows us to label raw GPS trajectories with vehicle classes, thereby enriching the data and enabling more comprehensive transportation studies such as derivation of vehicle class-specific origin-destination tables that can be used for planning. [ABSTRACT FROM AUTHOR]

In this paper, we present friendly jamming for secure location verification of vehicles in Intelligent Transportation Systems (ITS). In friendly jamming confidential information is obscured from eavesdroppers through the use of opportunistic jamming on the part of the parties engaged in communication. Secure location verification enables the infrastructure in an ITS to verify the position claims of a vehicle in an adversarial setting. We present a method for effectively using friendly jamming and distance bounding together to verify the location and velocity of vehicles, and detect any deviation of claims by a “prover” vehicle that is trying to prove its position and velocity to the infrastructure. The technique can be used to verify velocity and position claim of a vehicle, while keeping the feasible velocity and position lie by an attacker in a negligible range. Our analysis shows that in realistic traffic, friendly jamming is an improvement over comparable work, reducing position falsification by 86.66% to 92% for a typical attacker. The proposed technique also outperforms existing localization techniques for multiple colluding attackers, providing security against up to three colluding attackers, where colluders coordinate their movement and report information to help a single attacker achieve its goal. [ABSTRACT FROM AUTHOR]

As tourism and visitation increases to parks and protected areas worldwide, concerns regarding degradation of natural resources and visitor experiences also increase. As such, understanding visitor use patterns and spatial characteristics is important for better management of social and ecological resources. In many parks and protected areas, transportation systems (including park roads, parking lots, personal vehicles, public transit services, bicycle and/or pedestrian paths, and Intelligent Transportation Systems (ITS) that help deliver visitors to their destinations within parks) are a primary way for tourists and visitors to interact with and experience natural environments. Transportation systems designed with visitor use patterns in mind can both provide a positive experience for tourists as well as minimize impacts to natural resources within parks and protected areas. Our research used Global Positioning System (GPS)-based technology and a statistical classification procedure to examine spatial and temporal patterns of vehicular visitation in the Moose-Wilson corridor of Grand Teton National Park. While GPS-based technology has been used in parks and protected areas to study the spatial behavior of hikers and bikers, it has not been widely utilized to study the spatial behavior patterns of visitors travelling in vehicles to park destinations. Similarly, statistical classifications have been used to understand visitor behavior patterns in parks and protected areas, but are rarely applied to spatial data such as GPS tracks of visitor spatial movements. This classification based on spatial behavior resulted in the categorization of visitors into three types: Opportunistic Commuters, Wildlife/Scenery Viewers, and Hikers. These results expand existing literature suggesting that interpretable patterns of visitor spatial movements can result from classification studies using GPS-based collection of visitor behavior data. Understanding these patterns of visitor behavior is important for the management of social and natural resource conditions at park destinations as well as visitor experiences within the transportation corridor, as different types of visitors often exhibit different spatial behavior patterns. The classification and associated patterns of visitor behavior have important implications for managers of parks and protected area as different types of visitors may require different transportation and information needs in order to achieve their desired experience and minimize ecological impacts. [ABSTRACT FROM AUTHOR]