Heterogeneous Traffic Data Research Articles

Traffic Congestion Prediction Using Feature Series LSTM Neural Network and a New Congestion Index

Large and expanding cities suffer from a traffic congestion problem that harms the environment, travelers, and the economy. This paper aims to predict short term traffic congestion on a road section of expressway in Delhi city. For this purpose, we first propose a traffic congestion index based on traffic speed and flow. Clustering techniques and the Greenshield’s model were used for the derivation of the congestion index. Using this congestion index, congested time intervals of each day and each location of a weekday were identified. This study also introduces a feature series long short-term memory neural network (FSLSTMNN), which links a long short-term memory (LSTM) layer to each feature. It is trained using the many heterogeneous traffic features data collected in Delhi city for the next five minutes of traffic flow and speed prediction. FSLSTMNN achieved the good capability to learn feature series data. We also trained several traditional and deep-learning models using the same traffic data. The FSLSTMNN reduces mean absolute error 12.90% and 17.13%, respectively, in speed and traffic flow prediction compared to the second good-performance long short-term memory neural network (LSTMNN). Finally, traffic congestion is predicted classwise (light, medium, and congested) using the developed congestion index and traffic speed and flow predicted by the FSLSTMNN. Predicted results are consistent with the measured field data. Study results confirm that the developed congestion index and FSLSTMNN can be used successfully to predict traffic congestion.

Batch-based vehicle tracking in smart cities: A Data fusion and information integration approach

A data fusion and information integration (DFII-VT) framework is proposed to solve the batch-based vehicle matching/tracking problem using heterogeneous fixed-location and mobile sensing data available in smart cities. To make the model more realistic, traffic knowledge such as lane-choice decision, traffic merging, travel time and vehicle characteristics are calibrated using the historical dataset and then integrated into the model. The problem can be formulated as a combinatorial optimization model, and solved using a dynamic programming/Kuhn-Munkres algorithm-based two-step approach. By doing so, the proposed method can obtain individual travel times of the matched data pairs, and these can be directly used to estimate the corridor travel times of individual vehicles. The experimental results show that the fusion of mobile sensing data and fixed-location data yields significantly better results than using single-source data. Significant improvements in matching accuracy are obtained as more traffic information is integrated into the model. Therefore, the method may be considered a framework for integrating the manifold traffic information acquired within smart cities to obtain more accurate matching results and further used to optimize other fine-grained traffic applications, such as estimating vehicle trajectories along arterial corridors, estimating individual vehicle-based fuel consumption/emissions, and helping to infer real-time queuing processes at signalized intersections. The paper also studies some practical issues related to the use of heterogeneous traffic data, such as data errors, and detection failures.

TraGCAN: Trajectory Prediction of Heterogeneous Traffic Agents in IoV Systems

As a core component of the Internet of Vehicles, reasoning about the trajectory of pedestrians or vehicles in complex road conditions plays a critical role in autonomous driving and socially aware robotic navigation. Most existing methods do not adequately consider the effects of heterogeneous traffic agents. Toward this end, we propose the traffic trajectory prediction algorithm based on the convolutional attention network (TraGCAN) to predict the trajectories of heterogeneous traffic agents in dense traffic. The algorithm of the proposed method examines the behavior of different traffic agents in terms of both time and space dimensions to identify their movement patterns and interactions. We construct the spatial relationship of traffic agents as a graph structure and introduce a graph convolutional network to extract spatial interactions. In addition, we design a spatial attention mechanism to adaptively calculate weights for all spatial interactions to capture different influences from neighboring agents. To improve the accuracy of trajectory prediction, the algorithm considers the influence of the heterogeneous characteristics of traffic agents on their motion behaviors. We evaluated the performance of the proposed TraGCAN on heterogeneous traffic data sets, and the results demonstrate that the error of TraGCAN is reduced by 15% compared to existing methods.

Interoperable Internet of Things for Smart Transportation Systems in Circular Cities

This article discusses recent developments in the area of interoperable Internet of Things, using an example of delivering data-driven transportation services. Initial results obtained by exploiting heterogeneous traffic data streams reveal interesting traffic patterns for different vehicle types.

Multitask Hypergraph Convolutional Networks: A Heterogeneous Traffic Prediction Framework

Traffic prediction methods on a single-source data have achieved excellent results in recent years, especially the Graph Convolutional Networks (GCN) based models with spatio-temporal dependency. In reality, various modes of urban transportation operate simultaneously. They influence and complement each other in common space-time occasions, constituting the transportation system dynamically. Thus, traffic data from multiple sources is ostensibly heterogeneous, but internally correlated. The typical single data driven models are, however, not universally applicable for heterogeneous traffic data. To address this issue, we propose a Multi-task Hypergraph Convolutional Neural Network (MT-HGCN) for the multi-source traffic prediction problem. The framework consists of a main task and a related task. Both tasks are based on Hypergraph Convolutional Neural Networks (HGCN) and are devoted to two prediction problems. Furthermore, the tasks are bridged by a feature compress unit, which models the correlation and shares the latent feature to improve the performance of the main task. The node-level forecasting has been evaluated on historical datasets of Beijing to verify the effectiveness of the proposed method. Compared with the state-of-the-arts, the superior performance of the proposed method can be obtained.

An evolutionary computing‐based energy‐efficient solution for IoT‐enabled software‐defined sensor network architecture

SummarySoftware‐defined networking (SDN) has emerged as an evolving technique in wireless sensor networks (WSNs). SDN enables WSNs with programmable control to manage network functions dynamically and efficiently. In Internet of Things (IoT) applications, smart sensors suffer from the low battery issue, generally deployed in harsh network environments where regular recharge is not feasible. Moreover, integrating SDN with IoT‐enabled sensor network puts forward several challenges, for example, control nodes' selection, load balancing, and energy cost optimization while aggregating the collected data, focusing on heterogeneous traffic data. Thus, an energy‐efficient data collection technique via definite sensing control in two‐level IoT‐enabled software‐defined heterogeneous WSN (2SD‐HWSN) is formulated as an optimization problem, with transmission distance from smart sensors, residual energy of sensors, and load based on node density. The proposed algorithm is divided into two: set‐up and transmission phases. In the set‐up phase, the control server (CS) elects the best‐suited control nodes (CNs) and sets up a schedule for coordinating data transmission. Further, normal nodes join appropriate CNs based on distance and residual energy. This way, CNs form clusters and route sensed data during the transmission phase. Therefore, an alternative nature‐inspired algorithm, that is, grey wolf optimization (GWO), is hybridized with particle swarm optimization using a low‐level co‐evolutionary technique to improve its overall performance. This hybrid variant of GWO, known as HGWO‐BC, offers balanced clustering (BC) via novel fitness function design. An exhaustive simulation study is performed in different scenarios considering homogeneous and heterogeneous sensors. Comparative results show that the HGWO‐BC outperforms state‐of‐the‐arts concerning network lifetime, instability period, residual energy, throughput, and computational efforts.

Mining Actionable Patterns of Road Mobility From Heterogeneous Traffic Data Using Biclustering

The comprehensive access to road traffic patterns in the continuously growing urban areas is key to achieve a sustainable mobility. However, the inherent complexity of urban traffic poses many challenges to achieve this goal, including: i) the need to integrate heterogeneous views of road traffic (such as speed limits, jam size, delay, throughput) from available sources; ii) the complex spatiotemporal intricacies of geolocalized speed and loop counter data; iii) the need to mine congestion patterns robust to the inherent traffic variability and unexpected occurrence of events, taking also into consideration the varying degrees of congestion severity; and iv) the need to guarantee the statistical significance and interpretability of the target patterns. In the context of our work, a road traffic pattern is a recurrent congestion profile (w.r.t. speed limits, jam extent and flow) that can span multiple locations and time periods within a day. Biclustering, the discovery of coherent subspaces (local patterns) within real-valued data, has unique properties of interest, being positioned to unravel such traffic patterns, while satisfying the aforementioned challenges. Despite its relevance, the potentialities of applying biclustering in mobility domains remain unexplored. This work proposes a structured view on why, when and how to apply biclustering for mining traffic patterns of road mobility, a subject remaining largely unexplored up to date. Using the city of Lisbon as a guiding case, we illustrate the relevance of biclustering geolocalized speed data and loop counter data. The gathered results confirm the role of biclustering in comprehensively finding statistically significant and actionable spatiotemporal associations of road mobility.

Estimation for heterogeneous traffic using enhanced particle filters

This article explores the state estimation problem for heterogeneous traffic (vehicles with distinct driving behaviours) using particle filtering (PF) approaches. We consider three variations of PF to enhance estimation. The benchmark PF utilises a deterministic partial differential equation and a state-independent additive process noise. We first consider a parameter-adaptive PF variation that also allows model parameters to be adjusted. The second variation is a PF with spatially-correlated noise. The last variation combines parameter-adaptive and the spatially-correlated-noise approaches. We compare the four filters in numerical experiments that represent heterogeneous traffic scenarios and on real-world heterogeneous traffic data. The results show that the enhanced filters can achieve up to an 80% and 46% of accuracy improvement as compared to an open loop simulation without measurement correction, with the synthetic settings and with real traffic data, respectively. Moreover, the enhanced filters outperform the standard PF in all the traffic scenarios based on accuracy.

WCL: Client Selection in Federated Learning with a Combination of Model Weight Divergence and Client Training Loss for Internet Traffic Classification

Internet traffic classification (TC) is a critical technique in network management and is widely applied in various applications. In traditional TC problems, the edge devices need to send the raw traffic data to the server for centralized processing, which not only generates a lot of communication overhead but also leads to the privacy leakage and information security issues. Federated learning (FL) is a new distributed machine learning paradigm that allows multiple clients to train a global model collaboratively without raw traffic data sharing. The TC in a FL framework preserves the user privacy and data security by keeping the raw traffic data local. However, because of the different user behaviours and user preferences, traffic data heterogeneity emerges. The existing FL solutions introduce bias in model training by averaging the local model parameters from all heterogeneous clients, which degrades the classification accuracy of the learnt global classification model. To improve the classification accuracy in heterogeneous data environment, this paper proposes a novel client selection algorithm, namely, WCL, in federated paradigm based on a combination of model weight divergence and local model training loss. Extensive experiments on the public traffic dataset QUIC and ISCX have proved that the WCL algorithm obtains, compared to CMFL, superior performance in improving model accuracy and convergence speed on low heterogeneous traffic data and high heterogeneous traffic data, respectively.

Green Computing in Software Defined Social Internet of Vehicles

Social Internet of Vehicles (SIoV) is an evolving vehicular networking framework integrating the next generation smart devices with vehicular communications. Green computing and communication under disruptive vehicular environment is one of the challenging tasks for enabling SIoV. In this context, green traffic data dissemination in SIoV environments is modelled as an NP-hard problem focusing on heterogeneous traffic data, transmission distance from next generation smart devices and probabilistic delay in transmissions due to disruptive vehicular environment. An adopted meta-heuristic solution namely Two-Way Particle Swarm Optimization (TWPSO) is developed for the green traffic data dissemination problem in SIoV considering software defined vehicular network architecture. Extensive simulation experiments were performed to assess the performance of TWPSO as compared to the state-of-the-art techniques. The critical analysis of the comparative results attest the green computing oriented benefits of TWPSO under real SIoV environments.

Challenges faced in heterogeneous traffic data collection: a comparison of traffic data collection technologies

Traffic data are the fundamental inputs to traffic flow analysis and simulation studies, which facilitate decision making in the fi eld of traffic engineering. Hence, the accuracy of traffic data is of paramount importance. This study compares new technologies available for traffic data collection considering their accuracy and applicability in the Sri Lankan context. Traffic in Sri Lanka is of heterogeneous nature, as opposed to the homogeneous nature observed in most developed countries. Hence, collection of traffic data poses several challenges that affects its accuracy. Three techniques, the infrared driven TIRTL instrument, the video image processing-based TRAZER application and the traffic data collection method using the Google distance matrix application programming interface (API), with respect to their data collection accuracy are reviewed in this study. The fundamental macroscopic traffic data variables (fl ow and speed) were evaluated against control surveys. It was found that each technology has its strengths and weaknesses and needs to be used appropriately. The TIRTL instrument fared better on road sections on level terrain when the crossfall did not obstruct the infrared beams. Such occasions provided a rich set of microscopic traffic data. The TRAZER software delivered data up to a 100 % accuracy. However, this required the user to go through a lengthy postprocessing routine to extract the final set of traffic data. Google traffic data collection provides highly accurate results when estimating link speeds. This method is ideal for collection of bulk data with spatio-temporal variations and the process can be fully automated to reduce the human resource requirement.

On the modal shift from motorway to high-speed rail: evidence from Italy

Abstract The development of high-speed rail (HSR) has had a notable impact on modal market shares on the routes on which its services have been implemented. The aim of this study is to analyse whether the HSR expansion in Italy has led to a modal shift from motorway to HSR. We empirically test i) whether HSR openings adjacent to motorway sectors have reduced the total km travelled by light vehicles on these sectors during the period 2001–2017; and ii) whether this reduction has been persistent or even more evident after the opening of on-track competition between two HSR operators. To do so, we carried out a generalized difference-in-differences estimation, using a unique panel dataset that exploits the heterogeneous traffic data within all tolled motorway sectors in a quasi-experimental setting. Our findings reveal that neither HSR openings nor the opening of on-track competition led to a modal shift from motorway to HSR services, as the two transport modes are non-competing. Conversely, HSR expansion had a slightly positive impact on motorway traffic. The extent to which HSR demand could be the result of a modal shift from motorways is a relevant issue in any cost-benefit analysis of HSR investments.

Estimation of missing values in heterogeneous traffic data: Application of multimodal deep learning model

With the development of sensing technology, a large amount of heterogeneous traffic data can be collected. However, the raw data often contain corrupted or missing values, which need to be imputed to aid traffic condition monitoring and the assessment of the system performance. Several existing studies have reported imputation models used to impute the missing values, and most of these models aimed to capture the spatial or temporal dependencies. However, the dependencies of the heterogeneous data were ignored. To this end, we propose a multimodal deep learning model to enable heterogeneous traffic data imputation. The model involves the use of two parallel stacked autoencoders that can simultaneously consider the spatial and temporal dependencies. In addition, a latent feature fusion layer is developed to capture the dependencies of the heterogeneous traffic data. To train the proposed imputation model, a hierarchical training method is introduced. Using a real world dataset, the performance of the proposed model is evaluated and compared with that of several widely used temporal imputation models, spatial imputation models, and spatial–temporal imputation models. The experimental and evaluation results indicate that the values of the evaluation criteria of the proposed model are smaller, indicating a better performance. The results also show that the proposed model can accurately impute the continuously missing data. Furthermore, the sensitivity of the parameters used in the proposed deep multimodal deep learning model is investigated. This study clearly demonstrates the effectiveness of deep learning for heterogeneous traffic data synthesis and missing data imputation. The dependencies of the heterogeneous traffic data should be considered in future studies to improve the performance of the imputation model.

MF-TCPV: A Machine Learning and Fuzzy Comprehensive Evaluation-Based Framework for Traffic Congestion Prediction and Visualization

A framework for traffic congestion prediction and visualization based on machine learning and Fuzzy Comprehensive Evaluation named MF-TCPV is proposed in this paper. The framework uses DataX and DataV to implement the integration of multi-source heterogeneous traffic data and the visualization of congestion prediction results. A deep prediction model named LSTM-SPRVM based on deep learning algorithms, machine learning algorithms, and Spark parallelization technology for the prediction of traffic congestion features in the future is proposed. In MF-TCPV, traffic congestion is divided into six levels based on Fuzzy Comprehensive Evaluation and traffic congestion features such as average speed, road occupancy rate, and traffic flow density. MF-TCPV is validated based on the real data of Whitemud Drive in Canada. The experimental results demonstrate that MF-TCPV is capable of predicting the traffic congestion accurately and displaying prediction results visually. LSTM-SPRVM is better than other existing deep learning models in terms of prediction accuracy, and MF-TCPV can intuitively visualize the prediction results of traffic congestion.

Adaptive Rolling Smoothing With Heterogeneous Data for Traffic State Estimation and Prediction

Spatial–temporal traffic state estimation and the prediction of urban expressways is a vital component of traffic management and information systems. The adaptive smoothing method is one of the most frequently used approaches to estimate traffic states. However, the fixed filter parameters used in existing approaches sometimes fail to characterize traffic dynamics well. To better capture generation, propagation, and mitigation dynamics of traffic congestion, we propose an adaptive rolling smoothing (ARS) approach by dynamically tuning the filter parameters in a rolling horizon scheme for online applications. The fusion of heterogeneous traffic data combines aggregate traffic measurements (e.g., traffic flow rate, time occupancy, and speed collected by remote microwave sensors) and disaggregate information (e.g., timestamps of individual vehicles detected by license plate recognition cameras). A nonlinear traffic flow filter based on the virtual trajectory algorithm is established to reconstruct the spatial–temporal traffic state and estimate experienced travel times of individual vehicles. The results demonstrate the capability and effectiveness of the proposed ARS approach in the historical traffic state estimation and short-term traffic flow prediction. Complicated traffic states of weaving, merging, and diverging segments can be well distinguished by reconstructing time-space speed diagrams. The proposed approach can be extended to develop efficient missing data imputation algorithms and hierarchical control strategies for heterogeneously congested urban expressways.

Speed\u2013density functional relationship for heterogeneous traffic data: a statistical and theoretical investigation

This study is an attempt to establish a suitable speed–density functional relationship for heterogeneous traffic on urban arterials. The model must reproduce the traffic behaviour on traffic stream and satisfy all static and dynamic properties of speed–flow–density relationships. As a first attempt for Indian traffic condition, two behavioural parameters, namely the kinematic wave speed at jam (Cj) and a proposed saturation flow (λ), are estimated using empirical observations. The parameter Cj is estimated by developing a relationship between driver reaction time and vehicle position in the queue at the signalised intersection. Functional parameters are estimated using Levenberg–Marquardt algorithm implemented in the R statistical software. Numerical measures such as root mean squared error, average relative error and cumulative residual plots are used for assessing models fitness. We set out several static and dynamic properties of the flow–speed–density relationships to evaluate the models, and these properties equally hold good for both homogenous and heterogeneous traffic states. From the numerical analysis, it is found that very few models replicate empirical speed–density data traffic behaviour. However, none of the existing functional forms satisfy all the properties. To overcome the shortcomings, we proposed two new speed–density functional forms. The uniqueness of these models is that they satisfy both numerical accuracy and the properties of fundamental diagram. These new forms would certainly improve the modelling accuracy, especially in dynamic traffic studies when coupling with dynamic speed equations.

Exploring Human Mobility Patterns in Urban Scenarios: A Trajectory Data Perspective

Smart cities have been recognized as a promising research focus around the world. To realize smart cities, computation and utilization of big data are key factors. More specifically, exploring the patterns of human mobility based on large amounts of multi-source data plays an important role in analyzing the formation of social-economic phenomena in smart cities. However, our acquired knowledge is still very limited for smart cities. In this article, we propose an integrated computing method to rescale heterogeneous traffic trajectory data, which leverages MLE and BIC. Our analysis is based on two real datasets generated by subway smart card transactions and taxi GPS trajectories from Shanghai, China, which contain more than 451 million trading records by 14 subway lines and 34 billion GPS records by 13,695 taxis. Specifically, we quantitatively explore the patterns of human mobility on weekends and weekdays. Through logarithmic binning and data fitness, we calculate the Bayesian weights to select the best fitting distributions. In addition, we leverage three metrics to analyze the patterns of human mobility in two datasets: trip displacement, trip duration, and trip interval. We obtain several important human mobility patterns and discover quite a few interesting phenomena, which lay a solid foundation for future research.

SDN/NFV-based scalable mobile service integration for gradual network evolution

The coexistence of heterogeneous networks is becoming increasingly more common with the evolution towards 5G, increasing the complexity and cost of deploying and managing these networks. Network operators consequently need to find efficient ways of flexibly constructing and operating networks within a heterogeneous network environment in order to reduce operational expenditure (OPEX) and capital expenditure (CAPEX). One possible way to meet this requirement is the accommodation of previous legacy network services in the newly-deployed network while gradually reducing the reliance on physical legacy networks. Software-defined transitional network (SDTN) architecture is a strong candidate to achieve this. In the SDTN, the software-defined networking and the networkf function virtualization (SDN/NFV) plays a key role in integrating legacy network services into the newly-deploying network. This paper proposes an SDTN architecture and presents an example of its deployment in a mobile network evolution scenario in which 3G mobile services are integrated into a 4G network. The unified mobility management, in which 3G mobile services are integrated into a 4G access networks based on SDN/NFV is described. The network nodes for legacy mobility management services are virtualized at the control plane and accommodated into the SDTN architecture. As a result, the mobility management is interoperable between the virtualized legacy network service functions and the underlying 4G network. In order to support various heterogeneous services integration into the 4G mobile network, scalable edge switches are proposed for protocol mapping between incoming heterogeneous service traffic and 4G data. The proposed SDTN approach can be applied to the future deployment of 5G networks by accommodating various pre-existing legacy services in newly-deploying networks.

Highway Traffic Management System: Recent Trends and Future Challenges

Highway Traffic Management System (HTMS) is one of the important functional Systems of Intelligent Transportation Systems(ITS), which deals with data collection from roadside equipments, in order to provide more efficient services to highway users and to swiftly react in case of hazardous situations. More specifically, HTMS plays a vital role in road transportation systems, due to these reasons: a) High service capability in terms of both volume and speed, b) Exhibit limited controllability feature because of the nature of vehicular traffic, c) Recent technology developments enable new applications with the use of heterogeneous traffic data. This article provides a comprehensive summary of HTMS by providing a brief outline of existing literature and recent trends followed by the highlights of future research directions of this domain.

Freeway path travel time prediction based on heterogeneous traffic data through nonparametric model

ABSTRACTProviding reliable travel time prediction is very much needed for commuters for their upcoming trips to reduce travel time and relieve traffic congestion. This article proposes an integrated model for path and multi-step-ahead travel time prediction on freeways using both historical and real-time heterogeneous traffic and weather data. The model's performance is investigated in a case study under various traffic scenarios. Results indicate that the proposed model provides satisfactory prediction results in various performance tests. For practical purposes, general guidelines for selecting the model's parameter sets as well as the efficient size of historical data are also presented.