Traffic Detector Data Research Articles

Calibration of pedestrian ingress model based on CCTV surveillance data using machine learning methods.

Machine learning methods and agent-based models enable the optimization of the operation of high-capacity facilities. In this paper, we propose a method for automatically extracting and cleaning pedestrian traffic detector data for subsequent calibration of the ingress pedestrian model. The data was obtained from the waiting room traffic of a vaccination center. Walking speed distribution, the number of stops, the distribution of waiting times, and the locations of waiting points were extracted. Of the 9 machine learning algorithms, the random forest model achieved the highest accuracy in classifying valid data and noise. The proposed microscopic calibration allows for more accurate capacity assessment testing, procedural changes testing, and geometric modifications testing in parts of the facility adjacent to the calibrated parts. The results show that the proposed method achieves state-of-the-art performance on a violent-flows dataset. The proposed method has the potential to significantly improve the accuracy and efficiency of input model predictions and optimize the operation of high-capacity facilities.

Representing and Inferring Massive Network Traffic Condition: A Case Study in Nashville, Tennessee

Intelligent transportation systems (ITSs) usually require monitoring of massive road networks and gathering traffic data at a high spatial and temporal resolution. This leads to the accumulation of substantial data volumes, necessitating the development of more concise data representations. Approaches like principal component analysis (PCA), which operate within subspaces, can construct precise low-dimensional models. However, interpreting these models can be challenging, primarily because the principal components often encompass a multitude of links within the traffic network. To overcome this issue, this study presents a novel approach for representing and indexing network traffic conditions through weighted CUR matrix decomposition integrated with clustering analysis. The proposed approach selects a subset group of detectors from the original network to represent and index traffic condition through a matrix decomposition method, allowing for more efficient management and analysis. The proposed method is evaluated using traffic detector data from the city of Nashville, TN. The results demonstrate that the approach is effective in representing and indexing network traffic conditions, with high accuracy and efficiency. Overall, this study contributes to the field of network traffic monitoring by proposing a novel approach for representing massive traffic networks and exploring the effects of incorporating clustering into CUR decomposition. The proposed approach can help traffic analysts and practitioners to more efficiently manage and analyze traffic conditions, ultimately leading to more effective transportation systems.

Multivariate Poisson-Lognormal Models for Predicting Peak-Period Crash Frequency of Joint On-Ramp and Merge Segments on Freeways

Because of a growing crash occurrence in conflict areas, the ramp and merge segments on freeways are a concern for transportation researchers and practitioners. Therefore, short-term safety-performance functions (SPFs) have been proposed to predict crash frequency at a.m. and p.m. peak-period aggregation levels using microscopic traffic-detector data. The proposed short-term crash prediction models could achieve more accuracy and flexibility, give a better understanding of how safety evaluations change over time, and enable the taking of appropriate actions. This study contributes to the literature by using the multivariate Poisson-lognormal (MVPLN) method via an integrated nested Laplace approximation (INLA) approach to investigate the dependency and the correlation between two responses (on-ramp- and merge-related crash frequencies). Models are developed for total crashes (KABCO) and fatal and severe injury crashes (KAB), utilizing 70% of a total 239 and 238 for joint on-ramp and merge segments at a.m. and p.m. peaks, respectively, from three states of a freeway (i.e., Florida, Virginia, and Wisconsin). The traffic and specific geometric data (e.g., the number of lanes, ramp configurations, presence of weaving segment, and interchange connector type) for ramp and merge segments were used as independent variables. The significant variables were found to be the exposure parameters and various geometric feature variables for ramp and merge segments. Results of posterior means for the correlation coefficients between the ramp and merge crash frequencies indicate that a significant correlation exists between the two locations.

From Traffic Analysis to Real-Time Management: A Hazard-Based Modeling for Incident Durations Extracted Through Traffic Detector Data Anomaly Detection

The time and location of occurrence, among other characteristics of incidents, are needed to distribute traveler information and dispatch emergency service for traffic congestion and safety mitigation purposes. However, the corresponding clearance/recovery time is an important, yet under-studied, subject of research. Existing studies on incident duration modeling mainly utilize duration data gathered and provided by authorities through the efforts of individual agents. This type of manual data gathering may suffer coverage and consistency issues. In addition, the explanatory variables considered in these studies are static, that is, observed at the incident formation stage and assumed to be consistent over the incident periods. This setup, however, overlooks traffic flow dynamics; time-varying traffic variables during the incident episodes affect incidents clearance and disruption characteristics. In line with the above limitations, the objective of this study is to utilize hazard-based modeling to explain incident durations mined from traffic detector data while factoring in traffic flow dynamics. The 2014–2016 Virginia statewide traffic detector data were utilized to extract incident durations by implementing a machine-learning-based incident detection algorithm followed by a semi-parametric proportional hazard function. Such a function accounts for the time-varying traffic-descriptive covariates for incident duration modeling. The resulting monotonically decreasing baseline hazard reveals the snowball effect and the inertia effect of disruptions induced by incidents, which leads to the advice that incidents should be detected and mitigated within a 30 min duration. The normalized flow, density, and speed measures, and their temporal differences are all shown to be significant covariates scaling up or down the hazard.

Short-Term Traffic Speed Prediction for Multiple Road Segments

Short-term traffic prediction has been an essential part of real-time applications in modern transportation systems for the last few decades. Despite the recent progress in the voluminous models and data sources, many existing studies have focused on prediction for either a single or a few locations. In addition, the spatiotemporal dependency in the traffic data was narrowly accounted for. Therefore, this paper finds a new short-term traffic speed prediction algorithm that can efficiently cope with the complexity and immensity of the prediction process derived from the network size and amount of data in order to provide accurate predictions in real time. This algorithm consists of two modules: (a) principal component analysis (PCA) for data dimensionality reduction and feature selection, and (b) multichannel singular spectral analysis (MSSA) for multivariate time-series data prediction. A large amount of traffic data is efficiently compressed by PCA with high accuracy, then used as an input in the multivariate time-series analysis. The algorithm was compared with a vector autoregressive (VAR) model to predict traffic speeds five minutes ahead for a 21.3-mile-long highway segment, using the traffic detector data, and for 451-mile-long segment, using probe-based speed data in Tennessee. The tested algorithm is found to provide accurate predictions with a computation time of less than one second without training. Furthermore, the algorithm shows a better prediction performance under congested flow conditions, compared to VAR. This indicates that the tested algorithm is suitable for real-time prediction and scalable for a large network analysis.

Evaluating Real Time Ramp Meter Queue Length and Wait Time Estimation

Models to predict ramp meter queue length from traffic detector data are potentially useful tools in improving traffic operations and safety. Existing research, however, has been based on microscopic simulation or relied on extensive calibration of Kalman filter and related models to produce reliable queue length estimates. This research seeks to develop methodologies for improving and simplifying the calibration process of existing queue length models by applying loop detector data including volume, occupancy, and the metering rate data for ramp meters along I-15 in Utah. A conservation model and several variations of a Kalman filter model generated estimated queues that were compared to observed queue lengths in 60 s bins. A modified Kalman filter model and a new heuristic model derived from cluster analysis—the models that yielded the best results—provided queue length estimates that were generally within approximately eight vehicles of the observed queue length. Using the ramp metering rate, the queue length estimates were converted into wait times that were generally within approximately 30 s of the actual wait time, producing a viable method to predict wait time from up-to-the-minute traffic detection information with relatively little required calibration. The implementation of the ramp meter queue length and wait time estimation algorithms presented in this research will allow departments of transportation to better assess freeway and ramp conditions, which can then aid in reducing congestion throughout the freeway network.

Symbolic aggregate approximation based data fusion model for dangerous driving behavior detection

Detecting dangerous driving behavior is of great significance for reducing the occurrence of traffic accidents, and very challenging as it is affected by multiple factors. However, the existing methods are merely suitable for detecting dangerous driving behavior caused by single factor. Therefore, this paper proposes a novel Symbolic Aggregate approXimation based Data Fusion model (SAX-DF) to fuse multi-source data and applies it to accurately detect dangerous driving behavior. The proposed method considers multiple influencing factors at the same time and analyzes the related information between these factors. Further, a Positive Danger Mapping (PDM) algorithm is designed to unify the direction of multiple influencing factors. To verify the performance of SAX-DF, several experiments are carried out on traffic detector data and weather data. The experimental results illustrate that the proposed method can more effectively detect the dangerous driving behavior compared with the benchmark methods.

Spatiotemporal Attention-Based Graph Convolution Network for Segment-Level Traffic Prediction

Traffic prediction, as a core component of intelligent transportation systems (ITS), has been investigated thoroughly in the literature. Nevertheless, timely accurate traffic prediction still remains an open challenge due to the nonlinearities and complex patterns of traffic flows. In addition, most of the existing traffic prediction methods focus on grid-based computing problems (e.g., crowd in-out flow prediction) and point-based computing problems (e.g., traffic detector data prediction), ignoring the segment-based traffic prediction tasks. In this study, we propose an attention-based spatiotemporal graph attention network (AST-GAT) for segment-level traffic speed prediction. In particular, a multi-head graph attention block is designed to capture the spatial dependencies among road segments. Then, a component fusion block is built for speed, volume, and weather information integration. Finally, an attention-based Long short-term memory (LSTM) block is constructed for temporal dependency learning as well as segment-based speed prediction. Experiments on a real-world dataset from Highways England demonstrate that the proposed AST-GAT model outperforms the state-of-the-art baselines, providing an efficient tool for segment-based traffic prediction and therefore filling the gap between point-based and grid-based predictions.

Estimation of Intersection Turning Volume by Using Traffic Detector and Probe Data

Estimation of Intersection Turning Volume by Using Traffic Detector and Probe Data

Semi-supervised multiscale dual-encoding method for faulty traffic data detection

<abstract> <p>Inspired by the recent success of deep learning in multiscale information encoding, we introduce a variational autoencoder (VAE) based semi-supervised method for detection of faulty traffic data, which is cast as a classification problem. Continuous wavelet transform (CWT) is applied to the time series of traffic volume data to obtain rich features embodied in time-frequency representation, followed by a twin of VAE models to separately encode normal data and faulty data. The resulting multiscale dual encodings are concatenated and fed to an attention-based classifier, consisting of a self-attention module and a multilayer perceptron. For comparison, the proposed architecture is evaluated against five different encoding schemes, including (1) VAE with only normal data encoding, (2) VAE with only faulty data encoding, (3) VAE with both normal and faulty data encodings, but without attention module in the classifier, (4) siamese encoding, and (5) cross-vision transformer (CViT) encoding. The first four encoding schemes adopt the same convolutional neural network (CNN) architecture while the fifth encoding scheme follows the transformer architecture of CViT. Our experiments show that the proposed architecture with the dual encoding scheme, coupled with attention module, outperforms other encoding schemes and results in classification accuracy of 96.4%, precision of 95.5%, and recall of 97.7%.</p> </abstract>

Developing safety performance functions for freeways at different aggregation levels using multi-state microscopic traffic detector data

Safety Performance Functions (SPFs) have been widely used by researchers and practitioners to conduct roadway safety evaluation. Traditional SPFs are usually developed by using annual average daily traffic (AADT) along with geometric characteristics. However, the high level of aggregation may lead to a failure to capture the temporal variation in traffic characteristics (e.g., traffic volume and speed) and crash frequencies. In this study, SPFs at different aggregation levels were developed based on microscopic traffic detector data from California, Florida, and Virginia. More specifically, five aggregation levels were considered: (1) annual average weekday hourly traffic (AAWDHT), (2) annual average weekend hourly traffic (AAWEHT), (3) annual average weekday peak/off-peak traffic (AAWDPT), (4) annual average day of the week traffic (AADOWT), and (5) annual average daily traffic (AADT). Model estimation results showed that the segment length and volume, as exposure variables, are significant across all the aggregation levels. Average speed is significant with a negative coefficient, and the standard deviation of speed was found to be positively associated with the crash frequency. It is noteworthy that the operation of the high occupancy vehicle (HOV) lanes was found to have a positive effect on crash frequency across all the aggregation levels. The model results also showed that the AAWDPT and AADOWT models consistently performed better (the improvements range from 3.14%–16.20%) than the AADT-based SPF, which implies that the differences between the day of the week and peak/off-peak periods should be considered in the development of crash prediction models. The model transferability results indicated that the SPFs between Florida and Virginia are transferrable, while the models between California and the other two states are not transferrable.

Estimation of Queue Length at Signalized Intersections based on Electronic Police Data

Aiming at the problems of complicated traffic data detection environment in China and the larger calculation error of queue length at signalized intersections, a model of queue length estimation at signalized intersections based on electronic police data was studied. By analyzing the characteristics of travel time in the electronic police data, the jump position is extracted and the number of parking times is divided. Take the intersection of Zhongguan Road and Donghuan North Road in Ningbo as an example for simulation verification and analysis. Simulation results show that the accuracy rate of the model is over 83%, and the accuracy rate during peak hours is 90.8%, which has good adaptability.

Spatio-temporal traffic queue detection for uninterrupted flows

When traffic flow exceeds capacity because of demand fluctuations, crashes, work zones, and special events, a traffic queue is formed on a highway. Traffic queues cause potentially hazardous situations at the end of the queue where drivers unexpectedly face slowed or stopped traffic while approaching at high speed. Therefore, detecting a queue is vital for protecting it. This study presents a real-time spatio-temporal traffic queue detection algorithm that builds on traffic flow fundamentals combined with a statistical pattern recognition procedure. Using flow-density data, traffic flow phase is classified as either congested or uncongested flow in a probabilistic manner, based on Gaussian mixture models for each location in such a way that detects the traffic phase transitions. The proposed detection algorithm was applied to detect traffic queues using traffic detector data from Interstate 40 in Knoxville, Tennessee. The detection results show that the algorithm detects queues successfully by accounting for varying queueing conditions and different queue types. In addition, the proposed algorithm performed better than speed threshold methods in the literature.

Fault Data Detection of Traffic Detector Based on Wavelet Packet in the Residual Subspace Associated with PCA

To improve the accuracy and efficiency of fault data identification of traffic detectors is crucial in order to decrease the probability of unexpected failures of the intelligent transportation system (ITS). Since convolutional fault data recognition based on traffic flow three-parameter law has a poor capability for multiscale of fault data, PCA (principal component analysis) is adopted for traffic fault data identification. This paper proposes the fault data detection models based on the PCA model, MSPCA (multiscale principal component analysis) model and improved MSPCA model, respectively. In order to improve the recognition rate of traffic detectors’ fault data, the improved MSPCA model combines the wavelet packet energy analysis and PCA to achieve traffic detector data fault identification. On the basis of traditional MSPCA, wavelet packet multi-scale decomposition is used to get detailed information, and principal component analysis models are established on different scale matrices, and fault data are separated by wavelet packet energy difference. Through case analysis, the feasibility verification of traffic flow data identification method is carried out. The results show that the improved method proposed in this paper is effective for identifying traffic fault data.

Traffic state estimation on a two-dimensional network by a state-space model

This study proposes a state-space model that estimates traffic states over a two-dimensional network with alternative routes available by a data assimilation technique that fuses probe vehicle data with a traffic flow model. Although a number of studies propose traffic monitoring methods based on physical flow dynamics using sensing data such as probe vehicle and traffic detector data, they are basically limited to traffic monitoring along a simple road section. This study extends the analysis to a two-dimensional network, in which several alternative routes exist for each OD, with consideration of the route choice behaviours of users. Our proposed method employs sequential Bayesian filtering with a cell transmission model (CTM) for the flow model and probe vehicle data. From the probe vehicle data, not only cell densities but also diverging ratios are assumed to be measured and these measurements are assimilated into the flow model. The model validation in a hypothetical network reveals the potential of the model and discloses future issues.

Traffic State Estimation on a Two-Dimensional Network by a State-Space Model

Abstract This study proposes a state-space model that estimates traffic states over a two-dimensional network with alternative routes available by a data assimilation technique that fuses probe vehicle data with a traffic flow model. Although a number of studies propose traffic monitoring methods based on physical flow dynamics using sensing data such as probe vehicle and traffic detector data, they are basically limited to traffic monitoring along a simple road section. This study extends the analysis to a two-dimensional network, in which several alternative routes exist for each OD, with consideration of the route choice behaviours of users. Our proposed method employs sequential Bayesian filtering with a cell transmission model (CTM) for the flow model and probe vehicle data. From the probe vehicle data, not only cell densities but also diverging ratios are assumed to be measured and these measurements are assimilated into the flow model. The model validation in a hypothetical network reveals the potential of the model, and discloses future issues.

Revealing Recurrent Urban Congestion Evolution Patterns with Taxi Trajectories

Urban congestion can be classified into two types: Recurrent Congestion (RC) and Non-Recurrent Congestion (NRC). RC is more regular than NRC, having fixed and long-standing patterns. Mining urban recurrent congestion evolution patterns can assist with congestion cause analysis and the creation of alleviating strategies. Most existing methods for analyzing urban congestion patterns are based on traditional traffic detector data, which is inflexible and expensive. Additionally, prior research primarily focused on the microscopic model, which simulated congestion propagation based on theoretical models and hypothetical networks. As such, most previous models and methods are difficult to apply to real case scenarios. Therefore, we investigated recurrent congestion patterns by mining historical taxi trajectory data that were collected in Harbin, China. A three-step method is proposed to reveal urban recurrent congestion evolution patterns. Firstly, a grid-based congestion detection method is presented by calculating the change in taxi global positioning system (GPS) trajectory patterns. Secondly, a customized cluster algorithm is applied to measure the recurrent congestion area. Finally, a series of indicators are proposed to reflect RC evolution patterns. A case study was competed in the Harbin urban area to evaluate the main methods. Finally, RC cause analysis and alleviating strategy are discussed. The results study are expected to provide a better understanding of urban RC evolution patterns.

Missing data imputation for traffic flow speed using spatio-temporal cokriging

Modern transportation systems rely increasingly on the availability and accuracy of traffic detector data to monitor traffic operational conditions and assess system performance. Missing data, which occurs almost inevitably for a number of reasons, can lead to suboptimal operations and ineffective decisions if not remedied in a timely and systematic fashion through data imputation. A review of literature suggests that most traffic data imputation studies considered the temporal continuity of the data but often overlooked the spatial correlations that exist. Few of the studies explored the randomness of the patterns of the missing data. Therefore, this paper proposes two cokriging methods that exploit the existence of spatio-temporal dependency in traffic data and employ multiple data sources, each with independently missing data, to impute high-resolution traffic speed data under different data missing pattern scenarios. The two proposed cokriging methods, both using multiple independent data sources, were benchmarked against classic simple and ordinary kriging methods, which use only the primary data source. An array of testing scenarios were designed to test these methods under different missing rates (10–40% data loss) and different missing patterns (random in time and location, random only in location, and non-random blocks of missing data). The results suggest that using multiple data sources with the spatio-temporal simple cokriging method effectively improves the imputation accuracy if the missing data were clustered, or in blocks. On the other hand, if the missing data were randomly scattered in time and location, the classic ordinary or simple kriging method using only the primary data source can be more effective. Our study, which employs empirical traffic speed data from radar detectors and vehicle probes, demonstrates that the overall predictions of the kriging-based imputation approach are accurate and reliable for all combinations of missing patterns and missing rates investigated.

Rainfall‐integrated traffic speed prediction using deep learning method

Traffic information prediction is one of the most essential studies for traffic research, operation and management. The successful prediction of traffic speed is increasingly significant for the benefits of both road users and traffic authorities. However, accurate prediction is challenging, due to the stochastic feature of traffic flow and shallow model structure. Furthermore, environmental factors, such as rainfall influence, should also be incorporated to improve accuracy. Inspired by deep learning, this paper investigates the performance of deep belief network (DBN) and long short-term memory (LSTM) to conduct short-term traffic speed prediction with the consideration of rainfall impact as a non-traffic input. The deep learning models have the ability to learn complex features of traffic flow pattern under various rainfall conditions. To validate the performance of rainfall-integrated DBN and LSTM, the traffic detector data from an arterial in Beijing are utilised for model training and testing. The experiment results indicate that with the combination input of speed and additional rainfall data, deep learning models have better prediction accuracy over other existing models, and also yields improvements over the models without rainfall input. Furthermore, the LSTM can outperform the DBN to capture the time-series characteristics of traffic speed data.

Real-Time Simulation of Dynamic Traffic Flow with Traffic Data Assimilation Approach

The real-time traffic state estimation we propose uses a state-space model considering the variability of the fundamental diagram (FD) and sensing data. Serious congestion was caused by vehicle evacuation in many Sanriku coast cities following the great East Japan earthquake on March 11, 2011. Many of the vehicles in these congested queues were caught in the enormous tsunami after the earthquake [1]. Safe, efficient evacuation and rescue and restoration require that dynamic traffic states be monitored in real time especially in natural disasters. Variational theory (VT) based on kinematic wave theory is used for the system model, with probe vehicle and traffic detector data used to for measurement data. Our proposal agrees better with simulated benchmark traffic states than deterministic VT results do.