Large-scale Traffic Networks Research Articles

Integrating the Bi-dimensional-GSOM traffic flow interfaced models with a macroscopic emission model for very large-scale traffic networks

In this study, a coupled macroscopic traffic models and pollutant emission model for a very large-scale traffic networks is proposed. In order to overcome the missing data for large networks a continuum traffic modeling is developed aiming at simulating traffic of very large and dense networks. The present paper aims to merge both macroscopic traffic models (i. e. the Bi-dimensional model, BIDIM) with the generic second order model (GSOM) using real data. These two models will be associated with an emission model. The method consists of proposing an interface node model to ensure the exchange between the BIDIM and the GSOM traffic model. Then a vehicle emission model COPERT was adopted to provide the emission factors based on average speeds, in order to calculate macroscopic emissions for large networks. The dynamic of the developed simulation model reproduces the propagation of the congestion between both models and the evolution of emission pollutant in the networks. Applied to real networks, Marne La Vallée site and the A4 highway in France, the output results are very promising.

Change Detection in Partially Observed Large-Scale Traffic Network Data

Change Detection in Partially Observed Large-Scale Traffic Network Data

Demonstration-guided deep reinforcement learning for coordinated ramp metering and perimeter control in large scale networks

Effective traffic control methods have great potential in alleviating network congestion. Particularly, in an urban network consisting of heterogeneous roads (e.g., freeways and urban roads), how to integrate and coordinate control policies on different roads is a critical issue in large-scale networks. This study addresses this question from two aspects: modeling and control. From the modeling aspect, we formulate the hybrid traffic modeling in heterogeneous networks with the Asymmetric Cell Transmission Model (ACTM) for freeways and the generalized bathtub model for urban roads. For the control aspect, this study considers two representative control approaches: ramp metering for freeways and perimeter control for urban roads, and we aim to develop a deep reinforcement learning (DRL)-based coordinated control framework for large-scale networks. However, there are two significant challenges in the coordinated control in large-scale networks with DRL methods: non-stationary environment and large search space. To address both issues, we incorporate the demonstration to guide the DRL method for better convergence by introducing the concept of “teacher” and “student” models. The teacher models are traditional controllers that provide control demonstrations. For instance, ALINEA and Gating are two representative feedback controllers for ramp metering and perimeter control which can be “teacher” models. The student models are DRL methods, which learn from teachers and aim to surpass the teachers’ performance. Additionally, we develop a parallel training scheme to accelerate the proposed DRL method. To validate the proposed framework, we conduct two case studies in a small-scale network and a real-world large-scale traffic network in Hong Kong. Numerical results show that the proposed DRL method outperforms demonstrators as well as DRL methods, and the coordinated control is more effective than just controlling ramps or perimeters respectively. The research outcome reveals the great potential of combining traditional controllers with DRL for coordinated control in large-scale networks.

AI-powered intrusion detection in large-scale traffic networks based on flow sensing strategy and parallel deep analysis

Current intrusion detection systems, which rely on signature-based detection using rules derived from the inspection of past traffic flows and their signatures, are incapable of detecting new types of attacks. They also face challenges from large-scale traffic networks when deployed in the inline mode. This paper presents the SAID method combining signature-based and AI-powered deep analysis, called PAID, controllable by a flow sensing strategy for accelerating the intrusion detection in large-scale networks. PAID is designed as a parallel ensemble learning of DNN, XGB, and GBM to boost the deep analysis in terms of speed and quality. In order to enhance the performance of PAID, we also propose a method to augment the quality of the training dataset by compressing samples in the majority classes and generating more realistic samples in minority classes. Moreover, SAID includes a flow sensing strategy based on a configurable sampling cycle and window for analyzing the large-scale traffic network. Well-known datasets, such as CSE-CIC-IDS2018 and NSL-KDD, are used to demonstrate the performance of PAID. Our deep experiments prove that PAID obtains the same remarkable precision and F1-score of 99.97% and 99.69%, respectively. It also outperforms the other recent methods, and the typical PAID-based inspection time per flow (i.e., 4.89μs/flow) is fast enough to detect and prevent intrusions in real-time for large-scale networks.

A hybrid modelling framework for the estimation of dynamic origin–destination flows

The dynamic origin–destination flow estimation (DODE) problem requires scalable methods for large scale traffic networks and consistent techniques for capturing both uncongested and congested traffic conditions. Despite numerous efforts on incorporating multifold data sources and developing manifold mathematical models, the DODE problem remains a challenging problem in terms of both scalability and consistency. To fill this gap, we propose a novel hybrid DODE framework that integrates region-level (macro) and centroid-level (micro) traffic dynamics. The region-level traffic flows are described by the macroscopic fundamental diagram, while the centroid-level traffic flows are represented by the linear mapping of origin–destination flows onto link counts. This hybrid approach enables us to (i) incorporate region-level traffic measures into the problem, addressing scalability issues arising in large scale traffic networks, and (ii) capture non-linear behaviour of traffic in the regional context, enhancing consistency of the estimation results with respect to traffic conditions. The proposed methodology is experimented in a large-scale traffic network, which is benchmarked for DODE problems. The results indicate an outstanding performance of the hybrid DODE particularly in congested traffic conditions and highlight the effectiveness of aggregated (regional) traffic models in enhancing DODE methods with minimal computational burden.

Demand and state estimation for perimeter control in large-scale urban networks

State observability and demand estimation are two main issues in large-scale traffic networks which hinder real-world application of real-time control strategies. This study proposes a novel combined estimation and control framework (CECF) to develop perimeter control strategies based on macroscopic fundamental diagram (MFD). The proposed CECF is designed to operate with limited real-time traffic data and capture discrepancies in a priori demand estimates. The CECF is developed with a moving horizon estimator (MHE) that estimates traffic states, route choices and demand flows considering region accumulations and boundary flows observed from the network. The estimated traffic states are incorporated into a model predictive controller (MPC) scheme to derive the control decisions, which are then executed in the urban network. A novel accumulation-based MFD model is developed in this study to address the observability problem, which is incorporated into MHE and MPC schemes. The proposed CECF is implemented in a large-scale traffic network where several demand scenarios are tested. The results confirm the success of the CECF in overcoming observability issues and improving the performance of perimeter control strategies.

Hierarchical Cooperation and Load Balancing for Scalable Autonomous Vehicle Routing in Multi-Access Edge Computing Environment

When Connected Autonomous Vehicles (CAVs) request routing in a driving environment with Autonomous Intersection Management (AIM) systems, a routing planner collects the demands and optimizes the routes in a coordinated manner to reduce the overall travel times. In practice, however, the routing demands are massive, especially in a large-scale traffic network. As a result, the centralized routing planner fails to scale out to accommodate the growing requests, causing a severe scalability issue. This paper presents a holistic solution for scalable CAV routing by enabling hierarchical cooperation and load balancing in the Multi-Access Edge Computing (MEC) environment. The proposed system cooperatively plans CAV routes and dynamically balances loads in MECs to handle massive requests. According to the experiments, our system has a 15.68X higher routing capacity than the centralized routing system, and load balancing reduces 14.51% computation time of routing. The experiments show that our system is scalable for massive autonomous vehicle routing.

Submodularity of optimal sensor placement for traffic networks

The need for monitoring the state of a traffic network versus the costly installation and maintenance of roadside sensors constitutes the tough sensor placement problem in designing transportation networks. Placement problems naturally lie in the category of subset selection problems, which are known to be inherently combinatorial, and therefore, finding their exact solution is intractable for large problems. Due to this intractability, numerous heuristics have been proposed in the literature for approximately solving placement problems for traffic networks. Among these approaches, it has been observed that greedy algorithms normally outperform other heuristics. In this paper, we show the mathematics of why greedy algorithms are appropriate proxies for solving these subset selection problems; similar to placement problems for linear systems, placement problems for traffic networks also normally have a submodular structure. In this work, we analyze the problem of road sensor placement for transportation networks under different information structures available: when no vehicle routing information is available, when vehicles’ routings are known, and when it is necessary to maximize the number of origin–destination (O/D) traffic flows that are monitored with a set of sensors. We show that in all these cases, the placement problem has a submodular monotone structure. It is well known that the submodularity and monotonicity of discrete optimization problems can be leveraged to derive greedy algorithms that approximate the optimal solution. Consequently, our result is of great practical importance since by exploiting the submodularity and monotonicity of a problem, we show that it is possible to use polynomial-time greedy algorithms to approximate the combinatorial optimization problem with guaranteed optimality bounds for large problems, which are intractable to solve otherwise. Our results shed light upon the success of heuristic greedy algorithms that have been developed in some of the literature for solving placement problems at scale. To demonstrate the applicability of submodular optimization for solving placement problems, we first compare the performance of our polynomial-time approximation algorithm with the true optimum in an example traffic network which is small enough for finding the exact optimal solution with enumerating all possible subsets. Then, we investigate and validate our submodular approach in a case study involving a large-scale traffic network in Berkeley, California, where finding the exact optimal solution is intractable. Submodularity of the placement problem in these scenarios provides a powerful computational tool which can be further extended to other placement problem formulations that can become a reference for solving similar problems in the transportation literature.

Short-term traffic flow prediction model based on a shared weight gate recurrent unit neural network

Accurate traffic flow prediction is critical for enhancing traffic network operational efficiency. With the continuous expansion of traffic networks, providing reliable and efficient multi-step traffic flow prediction for large-scale traffic networks with a large number of sensors deployed has become a challenging issue. In this paper, we propose a multi-step many-to-many traffic prediction model for large-scale traffic networks, called spatio-temporal Shared GRU (STSGRU), which receive inputs from multiple sensors and provides predictions for all sensors simultaneously. First, we model the weekly pattern of traffic flow, using periodicity to explore long-term temporal features and provide smooth traffic flow to reduce the impact of data volatility. Second, different from existing models, we propose a shared weight mechanism to achieve many-to-many prediction without mapping traffic networks to images or graph structures. The proposed model strikes a delicate balance between complexity and accuracy. We validate the effectiveness of the proposed method on the Caltrans Performance Measurement System (PeMS) dataset. The results show that our model achieves similar prediction performance with advanced graph neural networks and has higher flexibility.

Toward Large-Scale Graph-Based Traffic Forecasting: A Data-Driven Network Partitioning Approach

Network partitioning is recognized as an effective auxiliary approach for solving transportation tasks on large-scale traffic networks in a domain-decomposition (DD) manner. Most of the existing related partitioning algorithms are explicitly designed to traffic management problems and merely focus on the implied topology of the networks. In this article, toward the practical problems that happened to traffic forecasting (TF) tasks, we propose a network-partitioning-based DD framework to improve graph convolutional network (GCN)-based predictors’ performance on large-scale transportation networks. Particularly, we devise a data-driven network-partitioning approach, namely, speed-matching-partitioning (SMP), which employs not only the topological features but also the traffic speed observations of traffic networks for partitioning. Additionally, we propose a data-parallel training strategy that feeds partitioned subnetworks into independent predictors for parallel training. The proposed approach is tested by comprehensive case studies on three real-world data sets to evaluate its effectiveness. The results indicate that the proposed approach can help improve GCN-based predictors’ accuracy and training efficiency on both small and relatively large traffic data sets. Furthermore, we investigate the model sensitivity to the selection of graph representations and framework parameters, and the learning efficiency of the data-parallel training strategy.

Reliability of the traffic network against cascading failures with individuals acting independently or collectively

The domino-like cascading failures may deteriorate traffic network reliability for both system and users. As the microscopic origin of this dynamic process, route choice behavior should be essential for the pattern of reliability against cascading failures, but the mechanism has not been systematically investigated. To uncover this, we propose a two-level cascading failure model to observe route choice patterns during the evolution of traffic jams: (i) the upper-level depicts how cascading failures spread out in network topologies, while (ii) the lower-level models what paths travelers may select in the trade-off between user equilibrium (UE) and system optimum (SO). Given nearly-one million OD pairs on large-scale traffic networks, the influence mechanisms of four factors on two reliability indicators are explored. We find that under certain perturbations, staying at the identified critical tolerance is the most profitable for investment yet unstable for reliability. Surprisingly, phase transition patterns are found in different topologies of traffic networks, where the lattice network is the most unreliable. Network reliability is found to increase by modifying travelers’ route choices from independent UE toward collective SO, whereas, the reliability is found to decrease if travelers have stochastic perception errors of travel time or have risk preferences to probabilistic travel time. Furthermore, if the overload failures can be recovered from the failed states, a traffic network may become more reliable. Our finding may help to improve reliability management of future transportation with multiple states between UE and SO.

Traffic Congestion Control Using Distributed Extremum Seeking and Filtered Feedback Linearization Control Approaches

This letter presents a hierarchical infrastructure-based control algorithm to manage mainstream traffic flow on freeways. At the upper level, a distributed Extremum-Seeking control approach is employed to determine the optimal density of vehicles in a congested cell. The local objective function is defined such that the average flow within the target cell is maximized to resolve the congestion, and the flow difference with its upstream cell is minimized to prevent back-propagating the congestion. At the lower level, a distributed Filtered Feedback Linearization controller is used to update the suggested velocity communicated to the vehicles so that the desired density determined by the upper level can be achieved in each cell. We adopted the METANET model to describe the aggregated dynamics of the traffic network. We tested the performance of these controllers via a MATLAB-VISSIM COM interface. The results demonstrate that the designed distributed controllers can achieve the desired closed-loop performance despite unknown disturbances in an uncertain large-scale traffic network.

Prediction of Emergency Mobility Under Diverse IoT Availability

INTRODUCTION: Prediction of emergency mobility needs to consider more scenarios as Internet of Things (IoT) develops at a high speed, which influences the quality and quantity of data, manageable resources and algorithms. OBJECTIVES: This work investigates differences in dynamic emergency mobility prediction when facing dynamic temporal IoT data with different quality and quantity considering diverse computing resources and algorithm availability. METHODS: A node construction scheme under a small range of traffic networks is adopted in this work, which can effectively convert the road to graph network structure data which has been proved to be feasible and used for the small-scale traffic network data here. Besides, two different datasets are formed using public large scale traffic network data. Representative widely used and proven algorithms from typical types of methods are selected respectively with different datasets to conduct experiments. RESULTS: The experimental results show that the graphed data and neural network algorithm can deal with the dynamic time series data with complex nodes and edges in a better way, while the non-neural network algorithm can predict the with a simple graph network structure. CONCLUSION: Our proposed graph construction with graph neural network improves dynamic emergency mobility prediction. The prediction should consider the scenarios of availability of computing resources, quantity and quality of data among other IoT features to improve the results. Later, automation and data enrichment should be improved.

Traffic Node Importance Evaluation Based on Clustering in Represented Transportation Networks

With the ability to acquire and process large-scale traffic big data, a cooperative intelligent transport system can be realized. Identifying important nodes in a traffic network contributes to better traffic control, which plays a more critical role in improving the traffic efficiency of a cooperative intelligent transport system. However, existing traffic node importance evaluation methods rely on manually designed metrics such as betweenness, degree, which may lead to biased results. Meanwhile, the traditional method of iteratively deleting nodes is unsuitable for a large-scale traffic network. In this paper, we propose a novel traffic node importance evaluation method based on clustering in represented transportation network. Specifically, the proposed method first construct a length-weighted network based on the geographic road network. Then, it learns the low-dimensional embeddings of nodes employing network representation learning. Finally, it clusters each nodes with a machine learning method and identifies the critical nodes through vehicle flow of different nodes. Experimental results on the real-word dataset show that our proposed method has excellent performance compared to baseline methods.

MS-Net: Multi-Source Spatio-Temporal Network for Traffic Flow Prediction

Predicting urban traffic flow is a challenging task, due to the complicated spatio-temporal dependencies on traffic networks. Urban traffic flow usually has both short-term neighboring and long-term periodic temporal dependencies. It is also noticed that the spatial correlations over different traffic nodes are both local and non-local. What’s more, the traffic flow is affected by various external factors. To capture the non-local spatial correlations, we propose a Dilated Attentional Graph Convolution (DAGC). The DAGC utilizes a dilated graph convolution kernel to expand the nodes’ receptive field and exploit multi-order neighborhood. Technically, the lower-order neighborhood corresponds to local spatial dependencies, while the higher-order neighborhood corresponds to non-local spatial dependencies between nodes. Based on DAGC, a Multi-Source Spatio-Temporal Network (MS-Net) is designed, which suffices to integrate long-range historical traffic data as well as multi-modal external information. MS-Net consists of four components: a spatial feature extraction module, a temporal feature fusion module, an external factors embedding module, and a multi-source data fusion module. Extensive experiments on three real traffic datasets demonstrates that the proposed model performs well on both the public transportation networks, road networks, and can handle large-scale traffic networks in particular the Beijing bus network which has more than 4,000 traffic nodes.

A quasi-contagion process modeling and characteristic analysis for real-world urban traffic network congestion patterns

The congestion propagation pattern in urban traffic networks has long been viewed as a complex question. Inspired by the fact that virus spread patterns are similar to traffic congestion propagation in some ways, this paper proposes a congestion propagation model (SIS-CP) to describe the congestion propagation patterns in large-scale traffic networks with very few parameters. In this SIS-CP model, the attribute of the real traffic network in the actual operation process is considered, a method based on Hilbert–Huang transform (HHT-based) is proposed to determine traffic states of links, and the congestion propagation patterns are analyzed. The real-world data of ‘DiDi Chuxing GAIA Open Dataset Initiative of Chengdu city from March to June is used to calibrate and validate the proposed models. The results show that the basic and limit value of congestion links’ ratio do exist in real traffic networks; the relationship between congestion links’ ratio and its derivative is conformed to the parabolic function; the HHT-based method can determine traffic states of links accurately; and the SIS-CP model can describe the congestion propagation patterns correctly.

A Method for Granular Traffic Data Imputation Based on PARATUCK2

Imputing missing data is a critical task in data-driven intelligent transportation systems. During recent decades there has been a considerable investment in developing various types of sensors and smart systems, including stationary devices (e.g., loop detectors) and floating vehicles equipped with global positioning system (GPS) trackers to collect large-scale traffic data. However, collected data may not include observations from all road segments in a traffic network for different reasons, including sensor failure, transmission error, and because GPS-equipped vehicles may not always travel through all road segments. The first step toward developing real-time traffic monitoring and disruption prediction models is to estimate missing values through a systematic data imputation process. Many of the existing data imputation methods are based on matrix completion techniques that utilize the inherent spatiotemporal characteristics of traffic data. However, these methods may not fully capture the clustered structure of the data. This paper addresses this issue by developing a novel data imputation method using PARATUCK2 decomposition. The proposed method captures both spatial and temporal information of traffic data and constructs a low-dimensional and clustered representation of traffic patterns. The identified spatiotemporal clusters are used to recover network traffic profiles and estimate missing values. The proposed method is implemented using traffic data in the road network of Manhattan in New York City. The performance of the proposed method is evaluated in comparison with two state-of-the-art benchmark methods. The outcomes indicate that the proposed method outperforms the existing state-of-the-art imputation methods in complex and large-scale traffic networks.

Scalable Traffic Signal Controls Using Fog-Cloud Based Multiagent Reinforcement Learning

Optimizing traffic signal control (TSC) at intersections continues to pose a challenging problem, particularly for large-scale traffic networks. It has been shown in past research that it is feasible to optimize the operations of individual TSC systems or a small collection of such systems. However, it has been computationally difficult to scale these solution approaches to large networks partly due to the curse of dimensionality that is encountered as the number of intersections increases. Fortunately, recent studies have recognized the potential of exploiting advancements in deep and reinforcement learning to address this problem, and some preliminary successes have been achieved in this regard. However, facilitating such intelligent solution approaches may require large amounts of infrastructure investments such as roadside units (RSUs) and drones, to ensure that connectivity is available across all intersections in the large network. This represents an investment that may be burdensome for the road agency. As such, this study builds on recent work to present a scalable TSC model that may reduce the number of enabling infrastructure that is required. This is achieved using graph attention networks (GATs) to serve as the neural network for deep reinforcement learning. GAT helps to maintain the graph topology of the traffic network while disregarding any irrelevant information. A case study is carried out to demonstrate the effectiveness of the proposed model, and the results show much promise. The overall research outcome suggests that by decomposing large networks using fog nodes, the proposed fog-based graphic RL (FG-RL) model can be easily applied to scale into larger traffic networks.

GAN and Multi-Agent DRL Based Decentralized Traffic Light Signal Control

Adaptive traffic light signal control (ATSC) is a promising paradigm for alleviating traffic congestion in intelligent transportation systems. Most of the existing methods require heavy traffic data exchange among neighboring intersections to achieve collaborative ATSC, which may not be supported by bandwidth-limited communication links in practice. In this article, we develop a communication-efficient decentralized ATSC framework for traffic networks with multiple intersections, where each intersection only exchanges traffic statistics with its neighboring intersections. In particular, the proposed framework consists of a generative adversarial network (GAN) based algorithm for traffic data recovery, and a multi-agent deep reinforcement learning (DRL) based decentralized ATSC algorithm for traffic efficiency enhancement. By adopting the value decomposition technique that establishes a nonlinear mapping from the local state-action values to the global reward, each intersection can independently determine its traffic light signal based on its local traffic data while achieving collaboration among neighboring intersections. Our proposed decentralized ATSC framework is scalable to large-scale traffic networks, and is also robust to traffic flow variations via interacting with the environment. Simulations show that our proposed algorithm can significantly reduce the vehicle travel time while maintaining high and stable traffic throughput.

A very large-scale traffic network modeling based on the integration of the Bi-dimensional and the GSOM Traffic Flow models

A continuum traffic modeling is proposed to be applicable in very large and dense networks, requiring a large volume of information difficult to recover and process. The aim of this paper is to merge both macroscopic models: bi-dimensional model (BIDIM) with the GSOM (Generic Second Order) model in order to overcome the missing data for large networks and to benefit from the advantages of each model. As well, the major advantage of the proposed approach is to reproduce a detailed description of traffic dynamics for the main arteries using the GSOM model. Indeed, the bi-dimensional model allows the simulation of areas where the measurements are missing or incomplete. The method consists in proposing an interface node model ensures the exchange between the BIDIM and the GSOM traffic models. The exchange between arteries of the GSOM model and cells edges of the bi-dimensional is assumed through the supply and demand concept.Moreover, the simulation is carried out using Godunov scheme and finite volume method. The dynamic of the developed simulation model reproduces the propagation of the congestion between both models. Applied to a synthetic network, the output results are very promising.