Traffic Data Sources Research Articles

Vehicle trajectory reconstruction for signalized intersections: A hybrid approach integrating Kalman Filtering and variational theory

Trajectory data provide a rich source of information on spatial and temporal speed fluctuations of individual vehicles. To assemble a complete picture of traffic operations, obtaining vehicle trajectories of the entire traffic flow is essential. The current techniques cannot fulfill this need sufficiently. To this end, we developed a hybrid approach for reconstructing vehicle trajectories at signalized intersections that accounts for queue dynamics. Specifically, an extended variational solution network was established and Kalman Filtering was applied to capture and reproduce the stochastic properties of queue boundary curves. The proposed method was demonstrated using both NGSIM data and simulation data. Results showed that the hybrid approach reproduces accurately the shockwave boundaries and trajectory reconstruction. By fusing different sources of traffic data, the proposed method does not rely on the common assumption of uniform arrival between adjacent probe vehicles, which is the main drive for estimation improvement.

Machine Learning Approach to Forecast Work Zone Mobility using Probe Vehicle Data

The aim of deploying intelligent transportation systems (ITS) is often to help engineers and operators identify traffic congestion. The future of ITS-based traffic management is the prediction of traffic conditions using ubiquitous data sources. There are currently well-developed prediction models for recurrent traffic congestion such as during peak hour. However, there is a need to predict traffic congestion resulting from non-recurring events such as highway lane closures. As agencies begin to understand the value of collecting work zone data, rich data sets will emerge consisting of historical work zone information. In the era of big data, rich mobility data sources are becoming available that enable the application of machine learning to predict mobility for work zones. The purpose of this study is to utilize historical lane closure information with supervised machine learning algorithms to forecast spatio-temporal mobility for future lane closures. Various traffic data sources were collected from 1,160 work zones on Michigan interstates between 2014 and 2017. This study uses probe vehicle data to retrieve a mobility profile for these historical observations, and uses these profiles to apply random forest, XGBoost, and artificial neural network (ANN) classification algorithms. The mobility prediction results showed that the ANN model outperformed the other models by reaching up to 85% accuracy. The objective of this research was to show that machine learning algorithms can be used to capture patterns for non-recurrent traffic congestion even when hourly traffic volume is not available.

Spatiotemporal variations in traffic activity and their influence on air pollution levels in communities near highways

Abstract Localized variations in traffic volume and speed can influence air pollutant emissions and corresponding concentrations in nearby communities, but most studies have utilized only aggregated traffic activity data. In this study, we compared the estimated influence of highway traffic activity on concentrations of primary oxides of nitrogen (NOx) and fine particulate matter (PM2.5) in communities near highways using a dispersion model informed by highly spatiotemporally-resolved variations of traffic volume and flow compared to the use of Annual Average Daily Traffic (AADT) data at a few locations. We used two sources of traffic activity data on 500 half-mile roadway segments on the five major highways in the Washington State Puget Sound during 2013. The first consisted of vehicle counts available every half-mile and 5 min; the second was traffic information (e.g., AADT) aggregated across the year and roadway network. Using the Motor Vehicle Emissions Simulator (MOVES) and the Research Line source dispersion model (RLINE), we modeled hourly concentrations of primary NOx and PM2.5 generated by highway traffic at nearly 4000 residences within 1 km of major highways. These concentrations were aggregated to daily and annual average concentrations, which were compared by input data source. At most locations, concentrations of primary NOx and PM2.5 modeled using the resolved traffic data had similar spatial and temporal distributions to concentrations predicted using the AADT data. However, several areas showed large differences. For example, 25% of residences within 150 m of a highway had concentrations that differed by more than 19% (8 ppb) for NOx and 32% (0.7 μg/m3) for PM2.5, and the AADT data consistently predicted lower concentrations than the resolved traffic data. Our findings indicate that temporal and spatial variation in traffic patterns can result in complex spatiotemporal variations of air pollutant concentrations that can be captured with the use of dispersion modeling with the appropriate inputs. The use of spatiotemporally resolved traffic activity data can improve exposure estimates and help reduce exposure measurement error in epidemiological studies, especially in communities near highly congested highways.

Analyzing Factors that Influence Expressway Traffic Crashes Based on Association Rules: Using the Shaoyang–Xinhuang Section of the Shanghai–Kunming Expressway as an Example

Analyzing the influencing factors in expressway traffic crashes is an important aspect of traffic safety analysis. This study examines detailed data from 217 crashes that took place from 2015 to 2016 on the Shaoyang–Xinhuang section of the Shanghai–Kunming Expressway. The data were recorded by the Expressway Administration of Hunan Province, China, which is the only officially available and reliable source of traffic accident data. Descriptive statistics for crash characteristics, vehicle conditions, and road environmental condition data are provided. This paper studies the characteristics of expressway traffic accidents and their influencing factors via the association rule data mining method. The 21 rules obtained from the association rules were analyzed, the accident characteristics of various types of vehicles were studied, and the causes of injury/fatal accidents in various situations were identified. The results of this study will contribute to the targeted enforcement of traffic regulations and the improvement of road facilities to reduce traffic casualties and promote road safety in other areas.

Road safety data collection systems in Iran: A comparison based on relevant organizations

PurposeVarious organizations and institutions are involved in road traffic injury (RTI) and crash registration such as police, forensic medicine organization, hospitals and emergency medical services. But there is a substantial uncertainty in interpreting the data, duplicated data collection and missing data in relation to RTI in most systems. This study aims to identify data sources for RTI surveillance in Iran and to explore traffic safety data source domains, data elements and detailed information by each data source. MethodsThis is a qualitative study which was conducted in 2017 in Iran. Data were collected employing semi-structured interviews with informants in road safety organizations in relation to traffic safety including Police, Ministry of Health and Medical Education as well as Forensic Medicine Organization and other authorities-in-charge. For completing the preliminary extraction information, the minimum data set was used and compared in each system. ResultsEight different organizations relevant to road traffic safety were identified. The main domain of data provided by each one consists of Emergency Medical System form, Police KAM114 form, Ministry of Transport and Road Administration, Red Crescent Organization/Disaster Management Information System, Ministry of Health and Medical Education, Forensic Medicine Organization, Insurance Company and Ministry of Justice. Each system has its own database, based upon its scope and mainly at crash and post-crash status and little on pre-crash circumstance. ConclusionAll current registry systems are not surveillance systems for RTI prevention. Huge data have been collected in various registry systems in Iran, but most of the collected variables are duplicated in each system. On the other hand, some variables like alcohol and substance abuse, child seat belt, helmet use in relation to RTI prevention are missed in all systems. Accordingly, it is a critical need to integrate and establish a comprehensive surveillance system, with focus on the goal of each system and collection of minimum data in each organization, which currently is underway.

Thresholds Based Image Extraction Schemes in Big Data Environment in Intelligent Traffic Management

Video traffic monitoring is an inexpensive and convenient source of traffic data. Traffic images processing are widely used to check traffic conditions and they can determine traffic control strategies in intelligent transportation systems (ITS). However, these traffic images always contain privacy-related data, such as vehicles registration numbers, human faces. Misuse of such data is a threat to the privacy of vehicles divers, passengers, pedestrians, etc. This paper proposes a thresholds-based images extraction solution for ITS. At first, a Faster Region Convolutional Neural Networks (RCNN) model is used to segment a traffic image into multi-regions with different importance levels; then, multi-threshold image extraction schemes are designed based on progressive secret image sharing schemes to extract images contain key traffic information, such as reg number, human faces, in which the region with higher importance level requires higher threshold for extraction. For different roles in ITS, they can extract images with different details, which can protect privacy and anonymity. The proposed methods provide a safe and intelligent way to extract images that can be used for further analysis in ITS.

A mixed integer programming formulation and scalable solution algorithms for traffic control coordination across multiple intersections based on vehicle space-time trajectories

Thanks to the development of mobile computing, novel traffic data sources are emerging as the promising building blocks for more effective traffic control strategies. It is expected that the vehicle space-time trajectories will become ubiquitously available in foreseeable future. Real-time trajectory data will provide full-spectrum pattern of traffic dynamics among multiple intersections. In this paper, we present a new traffic control representation for multiple intersections. A new multi-intersection phase (MI-phase) is proposed to represent safe vehicle movements across a few tightly connected intersections. All the intersections are also viewed as one integral “super intersection” within which vehicles move according to their planned paths. Through scheduling the sequence and durations of MI-phases over time, the vehicles will be crossing intersections with minimal delays. This approach can provide more flexibilities for traffic control coordination than the traditional Cycle-Split-Offset approach. A linear integer programming formulation is presented for joint optimization of vehicle space-time trajectories and traffic control. We also design a scalable optimization frame for real-world traffic control optimization, referred to as “Lagrangian decomposition with subproblem approximation” approaches. In this new framework, we construct the dynamic network loading based lower bound estimator (DNL-LBE) in which the relaxed constraints and sensitivity to the Lagrangian multiplier prices are explicitly considered while vehicular flows are being loaded. By doing so, the complex controlled dynamic network loading process can be represented through Lagrangian multipliers interfacing with the MI-phase optimization module (then solved by Dynamic Programming). This approach can facilitate price-based search heuristics to find high quality solutions for both vehicular space-time trajectories and traffic control plans without increasing the overall computing complexity. The efficiency of the proposed optimization framework is further improved through multiple advanced computing techniques. In the end, one demonstrative and one real-world example are provided to show the performance of the new approach.

Evaluating the Accuracy of Bluetooth-Based Travel Time on Arterial Roads: A Case Study of Perth, Western Australia

Bluetooth (BT) time-stamped media access control (MAC) address data have been used for traffic studies worldwide. Although Bluetooth (BT) technology has been widely recognised as an effective, low-cost traffic data source in freeway traffic contexts, it is still unclear whether BT technology can provide accurate travel time (TT) information in complex urban traffic environments. Therefore, this empirical study aims to systematically evaluate the accuracy of BT travel time estimates in urban arterial contexts. There are two major hurdles to deriving accurate TT information for arterial roads: the multiple detection problem and noise in BT estimates. To date, they have not been fully investigated, nor have well-accepted solutions been found. Using approximately two million records of BT time-stamped MAC address data from twenty weekdays, this study uses five different BT TT-matching methods to investigate and quantify the impact of multiple detection problems and the noise in BT TT estimates on the accuracy of average BT travel times. Our work shows that accurate Bluetooth-based travel time information on signalised arterial roads can be derived if an appropriate matching method can be selected to smooth out the remaining noise in the filtered travel time estimates. Overall, average-to-average and last-to-last matching methods are best for long (>1 km) and short (≤1 km) signalised arterial road segments, respectively. Furthermore, our results show that the differences between BT and ground truth average TTs or speeds are systematic, and adding a calibration is a pragmatic method to correct inaccurate BT average TTs or speeds. The results of this research can help researchers and road operators to better understand BT technology for TT analysis and consequently to optimise the deployment location and configuration of BT MAC address scanners.

A Method of Few-Shot Network Intrusion Detection Based on Meta-Learning Framework

Conventional intrusion detection systems based on supervised learning techniques require a large number of samples for training, while in some scenarios, such as zero-day attacks, security agencies can only intercept a limited number of shots of malicious samples. Therefore, there is a need for few-shot detection. In this paper, a detection method based on a meta-learning framework is proposed for this purpose. The proposed method can be used to distinguish and compare a pair of network traffic samples as a basic task of learning, including a normal unaffected sample and a malicious one. To accomplish this task, we design a deep neural network (DNN) named FC-Net, which mainly comprises two parts: feature extraction network and comparison network. FC-Net learns a pair of feature maps for classification from a pair of network traffic samples, then compares the obtained feature maps, and finally determines whether the pair of samples belongs to the same type. To evaluate the proposed detection method, we construct two datasets for few-shot network intrusion detection based on real network traffic data sources, using a specifically developed approach. The experimental results indicate that the proposed detection method is universal and is not limited to specific datasets or attack types. Training and testing on the same datasets demonstrate that the proposed method can achieve the average detection rate up to 98.88%. The outcome of training on one dataset and testing on the other one confirms that the proposed method can achieve better performance. In a few-shot scenario, malicious samples in an untrained dataset can be detected successfully, and the average detection rate is up to 99.62%.

Quality of floating car data (FCD) as a surrogate measure for urban arterial speed

Commercial floating car data (FCD) is being increasingly used as a traffic data source due to its lower cost despite concerns about its reliability. This paper focuses on the evaluation of FCD speed quality as a surrogate measure for arterial speed from different aspects. First, FCD speed is compared to video-based traffic data, collected from a specific urban road segment and assumed as ground truth in (a) descriptive evaluations, (b) speed estimation, and (c) level of service estimation. Regression analysis carried out to derive transformation function between two datasets showed a nonlinear relation with a high correlation coefficient of 0.82. Working with data along an urban corridor of 3.6 km also showed that despite some outliers, FCD was capable of detecting peak-hour queue formations as well as incident related ones. Use of transformation function on FCD speeds helped to increase its potential in urban traffic monitoring.

High spatial and temporal resolution vehicular emissions in south-east Brazil with traffic data from real-time GPS and travel demand models

Vehicular emissions are one of the most important source of pollution in urban centers, impacting air quality with a deleterious effect on human health and ecosystems. Air quality managers rely on emissions inventories to characterize pollution and sources. In this study we predicted vehicular emissions, using three sources of traffic data: 1) travel demand model outputs consisting of traffic simulations of light-duty vehicles, trucks, 2) and urban buses, and 3) a massive data set of real-time GPS coordinates of light-duty vehicles and trucks. The study area comprises the metropolitan areas of São Paulo, Santos, Vale de Paraíba, Sorocaba, and Campinas, which have a population of more than 30 million inhabitant. Once we generated hourly traffic flows, we used the Vehicular Emissions INventory Model (VEIN) to predict fuel consumption and emissions. Emissions using travel demand model for the metropolitan area of São Paulo are CO 177406 t/y, NOX 73554 t/y, NMHC 33999 t/y and PM2.5 2281 t/y. The emissions using GPS data were higher than using travel demand outputs, because GPS average speeds were lower, producing higher emission factors.

Measuring Traffic in Cities Through a Large-Scale Online Platform

Online real-time traffic data services could effectively deliver traffic information to people all over the world and provide large benefits to the society and research about cities. Yet, city-wide road network traffic data are often hard to come by on a large scale over a longer period of time. We collect, describe, and analyze traffic data for 45 cities from HERE, a major online real-time traffic information provider. We sampled the online platform for city traffic data every 5 min during 1 year, in total more than 5 million samples covering more than 300 thousand road segments. Our aim is to describe some of the practical issues surrounding the data that we experienced in working with this type of data source, as well as to explore the data patterns and see how this data source provides information to study traffic in cities. We focus on data availability to characterize how traffic information is available for different cities; it measures the share of road segments with real-time traffic information at a given time for a given city. We describe the patterns of real-time data availability, and evaluate methods to handle filling in missing speed data for road segments when real-time information was not available. We conduct a validation case study based on Swedish traffic sensor data and point out challenges for future validation. Our findings include (i) a case study of validating the HERE data against ground truth available for roads and lanes in a Swedish city, showing that real-time traffic data tends to follow dips in travel speed but miss instantaneous higher speed measured in some sensors, typically at times when there are fewer vehicles on the road; (ii) using time series clustering, we identify four clusters of cities with different types of measurement patterns; and (iii) a k-nearest neighbor-based method consistently outperforms other methods to fill in missing real-time traffic speeds. We illustrate how to work with this kind of traffic data source that is increasingly available to researchers, travellers, and city planners. Future work is needed to broaden the scope of validation, and to apply these methods to use online data for improving our knowledge of traffic in cities.

A deep learning approach to real-time parking occupancy prediction in transportation networks incorporating multiple spatio-temporal data sources

A deep learning model is adopted for predicting block-level parking occupancy in real time. The model leverages Graph-Convolutional Neural Networks (GCNN) to extract the spatial relations of traffic flow in large-scale networks, and utilizes Recurrent Neural Networks (RNN) with Long-Short Term Memory (LSTM) to capture the temporal features. In addition, the model is capable of taking multiple heterogeneously structured traffic data sources as input, such as parking meter transactions, traffic speed, and weather conditions. The model performance is evaluated through a case study in Pittsburgh downtown area. The GCNN-based model outperforms other baseline methods including multi-layer LSTM and LASSO with an average testing MAPE of 10.6% when predicting block-level parking occupancies 30 min in advance. The case study also shows that, in generally, the prediction model works better for business areas than for recreational locations. We found that incorporating traffic speed and weather information can significantly improve the prediction performance. Weather data is particularly useful for improving predicting accuracy in recreational areas.

Floating Car Data Technology

The limiting conditions of traffic in cities, together with the complex and dynamic traffic flows, require an efficient and systematic management and information provision for the traffic participants, with the goal to achieve better utilisation of traffic resources and preserve sustainable mobility. In that context, it is important to identify the traffic flow location features, which requires data and information. This paper presents the application of mobile vehicles for the collection of real time traffic flow data. Such data have become an important source of traffic data, since they can be collected in a simple and cost-efficient way, enabling higher coverage than the conventional approaches, despite the reliability issues. The term referring to that type of data collection, commonly used in scientific and professional literature is FCD (Floating Car Data) and “Probe vehicle”. The efficiency presentation of applying this extensive data source for retrieving necessary parameters and information related to the achievement of sustainable mobility is the final objective of this paper. A description of modern technologies that serve as a basis for probe vehicle data collection has been provided: a geographical information system (GIS), global navigation satellite system (GNSS) and related wireless communication. Within the key technologies review, the development possibilities of data collection by mobile sensors have also been presented.

Quasi-vehicle-trajectory-based real-time safety analysis for expressways

There have been plenty of segment-based real-time safety analyses for expressways or freeways. In these analyses, it was considered that the occurrence of a crash is because of segment conditions, including traffic and geometric. However, the movement of a vehicle is continuous from upstream to downstream. Thus, the crash occurrence of a vehicle might also be because of the traffic conditions along its trajectory. Under this idea, the study on the quasi-vehicle-trajectory-based real-time safety for expressways is proposed. It is hard to directly obtain the trajectory data of the vehicle involved in a crash. Thus, the crash-related vehicle’s trajectory is derived from other traffic data sources, and it is called quasi-trajectory. To find vehicles’ quasi-trajectories, the exact crash time was first identified by observing significant speed drops, and then, the vehicle locations for every one minute was distinguished by using the data from HERE, which provides space-mean speed information. The traffic information along the quasi-trajectories is from the Microwave Vehicle Detection System. A Bayesian matched-case-control logistic regression model was built to explore the impact of traffic parameters along the vehicle quasi-trajectory on real-time crash risk. The results showed that the segment speed difference and segment volume difference along the vehicle quasi-trajectory in the 0–4 min interval before crash occurrence all had a significant positive impact on crash risk. Meanwhile, the truck volume difference and average segment speed in the 0–1 min interval before crashes also have a positive impact on real-time crash likelihood. This study is new in real-time safety analysis and the findings have the potential to be applied to improve the safety of Connected and Autonomous Vehicles (CAV). A CAV continuously collects the traffic information along with its trajectory and could calculate the crash risk. When the hazardous traffic condition happens, actions can be taken by CAV to prevent the crash occurrence.

Expressway bottleneck pattern identification using traffic big data—The case of ring roads in Beijing, China

How to efficiently identify traffic bottlenecks in the expressway network and implement targeted measures is an important issue to mitigate traffic congestion. In recent years, with the proliferation of smartphones, smartphone-based navigation applications are receiving the worldwide popularity, which are being treated as an extremely rich source of traffic data. In this paper, we are devoted to proposing a novel approach to investigate traffic state and identify recurrent bottlenecks quickly and accurately from macroscopic network perspective, through fusing the data sources of fixed detectors and mobile navigation apps. First of all, this paper plots flow-speed fundamental diagram to derive critical speed, which is treated as the criterion to determine traffic state. Once the criterion has been established, typical bottlenecks in the entire expressway network can be efficiently identified through only utilizing smartphone-based probe speed data. Secondly, three pioneering indicators are put forward to quantify oversaturated traffic state and classify bottleneck patterns on urban expressway network. Applying this methodology, we take Beijing expressway network as a case to identify different patterns of bottlenecks, which are validated to comply with the reality. Moreover, the relationship between critical speed and the associated road segment features is explored and shows the possibility of predicting the critical speed even without flow data. As an end, some recommendations are proposed for improving different types of traffic bottlenecks.

Classification of Travel Modes Using Streaming GNSS Data

Abstract Over the last decade, smartphones became a valuable source of traffic data. GNSS data and other data from smartphone sensors can be successfully used in travel mode classification. Travel mode classification data are a significant source of information for various applications such as travel planning, urban road operations or user behavior understanding. Today, the availability of access to real-time data streams makes fast and real-time classification of travel modes possible. Because of different characteristics of data streams, the applied classification method has to be adjusted to the particular data stream. In this paper two classification methods, k Nearest Neighbors and Random Forest, are compared with emphasis on accuracy. First, they are applied for classification of travel modes using a static GNSS dataset, and afterward using streaming GNSS data. For the purpose of classification, characteristic distribution of velocity and acceleration for different travel modes is determined. Regarding streaming GNSS data, the influence of the window size on the classification accuracy is analyzed. Obtained results show that both classification methods can be successfully applied for the classification of travel modes.

Multi-source traffic data reconstruction using joint low-rank and fundamental diagram constraints

Accurate and complete traffic information plays an important role in traffic monitoring, traffic state estimation and prediction in all intelligent transportation systems. However, in practice, data provided from various types of traffic sensors may suffer from many issues, such as heterogeneity, missing data and bias. Therefore, extracting unbiased, accurate and complete data from multiple traffic data sources, referred to as multi-source traffic data reconstruction, is critical before data can be utilized. There are many existing methods for multi-source traffic data reconstruction, but two issues remain challenging and unaddressed: the effectiveness of reconstruction is fairly limited to attributes of available data, especially data loss ratio and reliability; the probability distribution of data assumed, which brings in bias prior to data analytics. This paper proposes a novel framework that integrates low-rank representation and fundamental traffic flow models into multi-source traffic data reconstruction. It makes full use of multi-source data, and can simultaneously impute missing data and eliminate outliers/errors. Multiple parameters in the traffic flow models are learned from large-scale spatio-temporal data, which ensures the extracted and imputed data are consistent with traffic flow physics. Experimental results on real-world traffic data demonstrate the superiority of this method over existing methods.

Traffic Big Data in the Development of Intelligent Transportation System

With the continuous development and advancement of computer technology, big data guarantees the establishment of an urban intelligent transportation system, a solid environmental basis to reform its application, and the construction of a deeply integrated data mechanism for big data-driven traffic management. This review paper briefly elaborates on the basic characteristics and sources of traffic big data as well as expound on the problems and application mechanisms of big data in intelligent transportation systems.

Enhancing Freeway Safety through Intervening in Traffic Flow Dynamics Based on Variable Speed Limit Control

New technologies and traffic data sources provide great potential to extend advanced strategies in freeway safety research. The High Definition Monitoring System (HDMS) data contribute comprehensive and precise individual vehicle information. This paper proposes an innovative Variable Speed Limit (VSL) based approach to manage crash risks by intervening in traffic flow dynamics on freeways using HDMS data. We first conducted an empirical analysis on real-time crash risk estimation using a binary logistic regression model. Then, intensive microscopic simulations based on AIMSUN were carried out to explore the effects of various intervention strategies with respect to a 3-lane freeway stretch in China. Different speed limits with distinct compliance rates under specified traffic conditions have been simulated. By taking into account the trade-off between safety benefits and delay in travel time, the speed limit strategies were optimized under various traffic conditions and the model with gradient feedback produces more satisfactory performance in controlling real-time crash risks. Last, the results were integrated into lane management strategies. This research can provide new ideas and methods to reveal the freeway crash risk evolution and active traffic management.