Macroscopic Traffic Simulation Research Articles

Machine Learning-Driven Calibration of Traffic Models Based on a Real-Time Video Analysis

Accurate traffic simulation models play a crucial role in developing intelligent transport systems that offer timely traffic information to users and efficient traffic management. However, calibrating these models to represent real-world traffic conditions accurately poses a significant challenge due to the dynamic nature of traffic flow and the limitations of traditional calibration methods. This article introduces a machine learning-based approach to calibrate macroscopic traffic simulation models using real-time traffic video stream data. The proposed method for creating and calibrating a traffic simulation model has significantly improved the statistical correspondence between the generated vehicle characteristics and real data about cars on the simulated road section. The correspondence has increased from 37% to 73%. Machine learning models trained on generated data and tested on real data show improved accuracy rates. Mean absolute error, mean square error, and mean absolute percentage error decreased by more than two orders of magnitude. The coefficient of determination has also increased, approaching 1. This method eliminates the need to deploy wireless sensor networks, which can reduce the cost of implementing intelligent transport systems.

Dynamic tradable credit scheme for multimodal urban networks

A Tradable Credit Scheme (TCS) is a demand management policy aiming for more sustainable travel behavior. The regulator defines the total credit cap and the credit distribution; it also determines the credit charges for each travel alternative at different times of the day, which modifies the perceived users’ costs. The credit price is determined by trading of credits between travelers. Defining the credit scheme at the urban level and estimating its impacts on user travel decisions and the network congestion dynamics is challenging. We propose a framework wherein travelers change their departure times and choose between solo car driving, Public Transportation (PT), and carpooling to complete their trips under a dynamic TCS, meaning the credit charge is time-dependent. A multimodal macroscopic traffic simulator based on a generalized bathtub model captures the congestion dynamics for the different transport modes. Additionally, we consider different values of time, trip lengths, and desired arrival times for the demand profile.The proposed TCS minimizes the total travel cost, (the sum of all travelers’ travel costs) and the carbon emissions by solving an optimization problem through an iterative method, including an inner loop that updates the users’ choices and credit price under the stochastic user equilibrium principle; and an outer loop that updates the credit charge profile. The methodology is implemented and applied to a realistic test case in Lyon (France). The dynamic TCS profiles result in 36% fewer carbon emissions than static TCS for a total travel cost reduction of 19%. Besides, 94% of the travelers benefit from the TCS as their user costs decrease in the case study. The final results are compared with a more advanced simulation framework, which is too computationally expensive to find the optimal TCS. The proposed TCS is effective with refined traffic dynamics representation.

A new reinforcement learning-based variable speed limit control approach to improve traffic efficiency against freeway jam waves

Conventional reinforcement learning (RL) models of variable speed limit (VSL) control systems (and traffic control systems in general) cannot be trained in real traffic process because new control actions are usually explored randomly, which may result in high costs (delays) due to exploration and learning. For this reason, existing RL-based VSL control approaches need a traffic simulator for training. However, the performance of those approaches are dependent on the accuracy of the simulators. This paper proposes a new RL-based VSL control approach to overcome the aforementioned problems. The proposed VSL control approach is designed to improve traffic efficiency by using VSLs against freeway jam waves. It applies an iterative training framework, where the optimal control policy is updated by exploring new control actions both online and offline in each iteration. The explored control actions are evaluated in real traffic process, thus it avoids that the RL model learns only from a traffic simulator. The proposed VSL control approach is tested using a macroscopic traffic simulation model to represent real world traffic flow dynamics. By comparing with existing VSL control approaches, the proposed approach is demonstrated to have advantages in the following two aspects: (i) it alleviates the impact of model mismatch, which occurs in both model-based VSL control approaches and existing RL-based VSL control approaches, via replacing knowledge from the models by knowledge from the real process, and (ii) it significantly reduces the exploration and learning costs compared to existing RL-based VSL control approaches.

Dynamic Hybrid Traffic Simulation Model at Merge and Diverge Sections at Highway

Traffic simulation is useful for testing many scenarios at a low cost and finding a good maintenance plan. However, detailed simulation in a wide area is costly. A dynamic hybrid traffic simulation model was proposed to achieve both precision and low computational cost simultaneously. In the dynamic hybrid model, a microscopic traffic simulation model is used where traffic jams occur, a macroscopic traffic simulation model is used in other areas, and these models are switched dynamically. The main target area of this model is a highway, but the model cannot handle merging and diverging sections currently. The authors are developing an improved dynamic hybrid model for these sections. In this paper, Cell Transmission Model is extended for merging and diverging sections, and a microscopic traffic simulation model was extended for merging sections.

Using Hybrid Clustering to Approximate Fastest Paths on Urban Networks

Estimating fastest paths on large networks is a crucial problem for dynamic route guidance systems. The present paper proposes a statistical approach for approximating fastest paths on urban networks. The traffic data used for conducting the statistical analysis is generated using a macroscopic traffic simulation software developed by us. The traffic data consists of the input flows, the arc states or the number of cars in the arcs and the paths joining the various origins and the destinations of the network. To find out the relationship between the input flows, arc states and the fastest paths of the network, we subject the traffic data to hybrid clustering. The hybrid clustering uses two methods namely k-means and Ward’s hierarchical agglomerative clustering. The strength of the relationship among the traffic variables was measured using canonical correlation analysis. The results of hybrid clustering are decision rules that provide fastest paths as a function of arc states and input flows. These decision rules are stored in a database for performing predictive route guidance. Whenever a driver arrives at the entry point of the network, the current arc states and input flows are matched against the database parameters. If agreement is found, then the database provides the fastest path to the driver using the corresponding decision rule. In case of disagreement, the database recommends the driver to choose the shortest path as the fastest path in order to reach the destination.

Macroscopic Traffic Simulation of Autonomous Vehicle Effects

The increasing worldwide demand on urban road transportation systems requires more restrictive measures and policies to reduce congestion, time delay and pollution. Autonomous vehicle mobility services, both shared and private, are possibly a good step towards a better road transportation future. This article aims to study the expected impact of private autonomous vehicles on road traffic parameters from a macroscopic level. The proposed methodology focuses on finding the different effects of different combinations of autonomous vehicle penetration and Passenger Car Units (PCU) on the chosen road traffic model. Four parameters are studied: traveled daily kilometers, daily hours, total daily delay and average network speed. The analysis improves the four parameters differently by implementing autonomous vehicles. The parameter total delay has the most significant reduction. Finally, several mathematical models are developed for the percentage of improvement for each chosen parameter.

Impact of Information on Topology-Induced Traffic Oscillations

Previous studies have shown traffic oscillations can be induced by special network topology. In the simplest case, a network of two intersections connected by two parallel roads would produce oscillatory traffic, when the split of drivers between the two roads falls into certain range. To understand how traffic information may affect such oscillations, a subset of drivers is allowed to be “reactive” in this study; that is, their route choice varies according to information about prevailing traffic conditions on the roads. We show that, depending on the ratio of reactive drivers, the system displays six new decaying, periodic oscillatory, or stable patterns. All solutions are obtained analytically in closed form and validated by macroscopic traffic simulation. Of all the solutions discovered, only one both is stable and fully utilizes the road space between the two intersections, and hence it is more desirable than the other solutions. The findings reveal the link between information provision and topology-induced oscillations, which may help practitioners design strategies that contribute to mitigating the adverse impact of such oscillations.

A generic work zone evaluation tool driven by a macroscopic traffic simulation model

A generic work zone evaluation tool driven by a macroscopic traffic simulation model

Design of reinforcement learning for perimeter control using network transmission model based macroscopic traffic simulation

Perimeter control is an emerging alternative for traffic signal control, which regulates the traffic flows on the periphery of a road network. Some model-based approaches have been suggested earlier for the optimization of perimeter control based on macroscopic fundamental diagrams (MFDs). However, there are several limitations when considering their application to a large-scale urban area because the model-based approaches may not be scalable to multiple regions and inappropriate for handling various effects caused by the shape change of MFDs. Therefore, we propose a model-free and data-driven approach that combines reinforcement learning (RL) with the macroscopic traffic simulation based on the recently developed network transmission model. First, we design four perimeter control models with different macroscopic traffic variables and parametrizations. Then, we validate the proposed models by evaluating their performances with the test demand scenarios at different levels. The validation results show that the model containing travel demand information adapts to a new demand scenario better than the model containing only density-related factors.

A Generic Work Zone Evaluation Tool Driven by Macroscopic Traffic Simulation Model

This study proposed a generic work zone evaluation tool driven by a dynamic macroscopic traffic model. The tool can investigate the effects of a work zone on freeway conditions. The classical freeway traffic network model METANET was extended to better capture typical work zone characteristics in terms of speed and capacity reductions through modifying its speed-density relationship functions. The tool can present profiles of traffic state variables, system-level performance indices, and annotated traffic conditions maps. It was used to develop a prototype computer-aided decision support system and applied in two simulation-based case studies. Case studies show that it is effective and useful for understanding complicated work zone impacts and can aid in better design and assessment of work zone plans. It can be integrated in a dynamic transportation management system for better communication between road operator and travellers.

Local and coordinated ramp metering within the unifying framework of an adaptive control scheme

This study develops and proposes a unifying control scheme to address the ramp and mainline metering problems arising in freeways. A distinguishing novelty of the proposed approach is that it can inherently be applied both at local and coordinated levels. The proposed strategy is based on a recently developed nonlinear adaptive control scheme, consisting of a nominal feedback law in conjunction with a nonlinear observer, which aims to estimate some unknown system variables. The present study provides insights into and demonstrates the properties and the performance of the proposed methodology, which is tested for realistic traffic scenarios using a macroscopic traffic flow simulator as a surrogate for potential field application. Comparison tests have been performed with other control strategies proposed in the literature and employed already in the field.

Comparing the lane-changing behaviour of Iran and the United States

The unique geometric features, the high-speed limit and the presence of multiple lanes in highways allow them to support large volumes of traffic. However, driving behaviours such as lane-change manoeuvres reduce the functional capacity of freeways, thereby increasing the travel time and decreasing the safety of travel. Lane changes can be generally divided into mandatory and discretionary manoeuvres. The aim of the discretionary lane change is to improve driving conditions while mandatory lane change is forced on the driver following the target path. The aim of this study was to provide a count-regression model to predict the number of discretionary lane changes in freeways. The study used data pertaining to the interstate highway (I-80) in the United States and Modarres highway in Tehran, Iran, to develop linear-regression, Poisson-regression and negative-binomial models. The performance of these models and lane-changing behaviours in Iran and the United States were also compared. The developed model can be utilised in macroscopic traffic simulation software and for the determination of the level of service. According to the results, the number of vehicles in the current lane has the greatest effects on the quality of results and a comparison between Iran and the United States showed interesting outputs.

Traffic Modeling for Wildland–Urban Interface Fire Evacuation

Several traffic modeling tools are currently available for evacuation planning and real-time decision support during emergencies. This paper reviews potential traffic-modeling approaches in the context of wildland-urban interface (WUI) fire-evacuation applications. Existing modeling approaches and features are evaluated pertaining to fire-related, spatial, and demographic factors; intended application (planning or decision support); and temporal issues. This systematic review shows the importance of the following modeling approaches: dynamic modeling structures, considering behavioral variability and route choice; activity-based models for short-notice evacuation planning; and macroscopic traffic simulation for real-time evacuation management. Subsequently, the modeling features of 22 traffic models and applications currently available in practice and the literature are reviewed and matched with the benchmark features identified for WUI fire applications. Based on this review analysis, recommendations are made for developing traffic models specifically applicable to WUI fire evacuation, including possible integrations with wildfire and pedestrian models.

A macroscopic first-order traffic simulation with a modified dynamic node approach

ABSTRACTIn this work, we present the different steps in the modeling and implementation of a first-order macroscopic road traffic simulation tool, using Lighthill–Whitham–Richards approach with a Godunov scheme (cell transmission model) and bounded acceleration. The origin–destination traffic movement is based on the definition of partial streams according to their chosen paths or ‘routes’ from an origin to its destination. The resolution of the traffic conflicts at the nodes level is achieved using a modified dynamic node model. The dynamic node model proposed in this work introduces a new supply function which follows from the extension of the Highway Capacity Manual (2010) model for use in dynamic and instantaneous regime. The method is then applied to reconstruct the traffic at a roundabout in the city of Sousse, Tunisia, known to be over-jammed at rush hour. The results are satisfactory and can be generalized to other intersection types with different priority levels.

Traffic Simulation in Urban Networks Using Stochastic Cell Transmission Model

The stochastic cell transmission model (SCTM) is a macroscopic traffic simulation model of high accuracy. One of its advantages is that it can represent the uncertainty of the traffic states and the changing travel demand and supply conditions. However, it has been applied to freeways and simple networks having only one origin‐destination pair. In this article, we propose a modified stochastic cell transmission model (M‐SCTM) that applies the conventional SCTM to urban networks. In M‐SCTM, we introduce vehicle agents as well as their route choice behavior on an urban network, which is more applicable to complex urban networks. Additionally, M‐SCTM was applied to networks in which the turning ratio is not priorly set through a route search algorithm. The results show that M‐SCTM can conduct simulations with as much accuracy as SCTM. Furthermore, we verified the appropriate reproducibility of the simulations based on M‐SCTM and compared the estimated value and the measured value in terms of the travel time of each vehicle.

Traffic and emissions impact of congestion charging in the central Beijing urban area: A simulation analysis

Congestion charging is being considered as a potential measure to address the issue of substantially increased traffic congestion and vehicle emissions in Beijing. This study assessed the impact of congestion charging on traffic and emissions in Beijing using macroscopic traffic simulation and vehicle emissions calculation. Multiple testing scenarios were developed with assumptions in different charging zone sizes, public transit service levels and charging methods. Our analysis results showed that congestion charging in Beijing may increase public transit use by approximately 13%, potentially reduce CO and HC emissions by 60–70%, and reduce NOx emissions by 35–45% within the charging zone. However, congestion charging may also result in increased travel activities and emissions outside of the charging zone and a slight increase in emissions for the entire urban area. The size of charging zone, charging method, and charging rate are key factors that directly influence the impact of congestion charging; improved public transit service needs to be considered as a complementary approach with congestion charging. This study is used by Beijing Transportation Environment and Energy Center (BTEC) as reference to support the development of Beijing’s congestion charging policy and regulation.

Optimal Dynamic Tolls for Managed Lanes

This paper presents a real-time simulation-based control framework to determine dynamic toll rates to optimize an operator's objective subject to various operational and contractual constraints, such as smooth toll rate changes and maintenance of prescribed levels of service on the toll lane. The toll-setting system incorporates models to predict both the vehicle arrival process upstream of the toll lane facility and drivers’ choice whether to use the toll lanes as a function of the toll rate and travel times presented to drivers within the information system. A macroscopic traffic simulation model is used to predict the flow conditions within the prediction horizon. The travel times provided to users as information and the travel times predicted by the traffic flow model are iterated until consistency between them is obtained. The whole process is embedded within an optimization algorithm that sets tolls to optimize a given objective function. Several case studies demonstrate the use of this framework and its potential to provide useful toll settings.

Robust calibration of macroscopic traffic simulation models using stochastic collocation

The predictions of a well-calibrated traffic simulation model are much more valid if made for various conditions. Variation in traffic can arise due to many factors such as time of day, work zones and weather. Calibration of traffic simulation models for traffic conditions requires larger datasets to capture the stochasticity in traffic conditions. In this study we use datasets spanning large time periods to incorporate variability in traffic flow, speed for various time periods. However, large data poses a challenge in terms of computational effort. With the increase in number of stochastic factors, the numerical methods suffer from the curse of dimensionality. In this study, we propose a novel methodology to address the computational complexity due to the need for the calibration of simulation models under highly stochastic traffic conditions. This methodology is based on sparse grid stochastic collocation, which, treats each stochastic factor as a different dimension and uses a limited number of points where simulation and calibration are performed. A computationally efficient interpolant is constructed to generate the full distribution of the simulated flow output. We use real-world examples to calibrate for different times of day and conditions and show that this methodology is much more efficient that the traditional Monte Carlo-type sampling. We validate the model using a hold out dataset and also show the drawback of using limited data for the calibration of a macroscopic simulation model. We also discuss the drawbacks of the predictive ability of a single calibrated model for all the conditions.

Simulation-based Optimization of HOT Lane Tolls

This paper presents a real-time simulation-based control framework to determine dynamic toll rates in order to optimize an operator's objective, subject to various operational and contractual constraints, such as smooth toll rate changes and maintaining prescribed levels of service on the toll lane. The toll-setting system incorporates models to predict the vehicle arrival process upstream of the toll lane facility and drivers’ choice whether or not to use the toll lanes as a function of the toll rate and travel times presented to drivers within the information system. A macroscopic traffic simulation model is used to predict the flow conditions within the prediction horizon. The travel times provided to users as information and the ones predicted by the traffic flow model are iterated until consistency between them is obtained. The whole process is embedded within an optimization algorithm that sets tolls that optimize a given objective function. Several case studies demonstrate the use of this framework and its potential to provide useful toll settings.

Traffic Flow Prediction for Road Transportation Networks With Limited Traffic Data

Obtaining accurate information about current and near-term future traffic flows of all links in a traffic network has a wide range of applications, including traffic forecasting, vehicle navigation devices, vehicle routing, and congestion management. A major problem in getting traffic flow information in real time is that the vast majority of links is not equipped with traffic sensors. Another problem is that factors affecting traffic flows, such as accidents, public events, and road closures, are often unforeseen, suggesting that traffic flow forecasting is a challenging task. In this paper, we first use a dynamic traffic simulator to generate flows in all links using available traffic information, estimated demand, and historical traffic data available from links equipped with sensors. We implement an optimization methodology to adjust the origin-to-destination matrices driving the simulator. We then use the real-time and estimated traffic data to predict the traffic flows on each link up to 30 min ahead. The prediction algorithm is based on an autoregressive model that adapts itself to unpredictable events. As a case study, we predict the flows of a traffic network in San Francisco, CA, USA, using a macroscopic traffic flow simulator. We use Monte Carlo simulations to evaluate our methodology. Our simulations demonstrate the accuracy of the proposed approach. The traffic flow prediction errors vary from an average of 2% for 5-min prediction windows to 12% for 30-min windows even in the presence of unpredictable events.