Showing papers on "Traffic simulation published in 2020"

Errata for “Cooperative Eco-Driving at Signalized Intersections in a Partially Connected and Automated Vehicle Environment”

Abstract: The emergence of connected and automated vehicle (CAV) technology has the potential to bring a number of benefits to our existing transportation systems. Specifically, when CAVs travel along an arterial corridor with signalized intersections, they can not only be driven automatically using pre-designed control models but can also communicate with other CAVs and the roadside infrastructure. In this paper, we describe a cooperative eco-driving (CED) system targeted for signalized corridors, focusing on how the penetration rate of CAVs affects the energy efficiency of the traffic network. In particular, we propose a role transition protocol for CAVs to switch between a leader and following vehicles in a string. Longitudinal control models are developed for conventional vehicles in the network and for different CAVs based on their roles and distances to intersections. A microscopic traffic simulation evaluation has been conducted using PTV VISSIM with realistic traffic data collected for the City of Riverside, CA, USA. The effects on traffic mobility are evaluated, and the environmental benefits are analyzed by the U.S. Environmental Protection Agency’s MOtor Vehicle Emission Simulator (MOVES) model. The simulation results indicate that the energy consumption and pollutant emissions of the proposed system decrease, as the penetration rate of CAVs increases. Specifically, more than 7% reduction on energy consumption and up to 59% reduction on pollutant emission can be achieved when all vehicles in the proposed system are CAVs.

A Survey on Visual Traffic Simulation: Models, Evaluations, and Applications in Autonomous Driving

Abstract: Virtualized traffic via various simulation models and real‐world traffic data are promising approaches to reconstruct detailed traffic flows. A variety of applications can benefit from the virtual traffic, including, but not limited to, video games, virtual reality, traffic engineering and autonomous driving. In this survey, we provide a comprehensive review on the state‐of‐the‐art techniques for traffic simulation and animation. We start with a discussion on three classes of traffic simulation models applied at different levels of detail. Then, we introduce various data‐driven animation techniques, including existing data collection methods, and the validation and evaluation of simulated traffic flows. Next, we discuss how traffic simulations can benefit the training and testing of autonomous vehicles. Finally, we discuss the current states of traffic simulation and animation and suggest future research directions.

The impact of autonomous vehicles on urban traffic network capacity: an experimental analysis by microscopic traffic simulation

Abstract: Urban commuters have been suffering from traffic congestion for a long time. In order to avoid or mitigate the congestion effect, it is significant to know how the introduction of autonomous vehicl...

Analysing urban traffic volumes and mapping noise emissions in Rome (Italy) in the context of containment measures for the COVID-19 disease

Abstract: This study presents the result of a traffic simulation analysis based on Floating Car Data and a noise emission assessment to show the impact of mobility restriction for COVID-19 containment on urban vehicular traffic and road noise pollution on the road network of Rome, Italy The adoption of strong and severe measures to contain the spreading of Coronavirus during March-April 2020 generated a significant reduction in private vehicle trips in the city of Rome (-646% during the lockdown) Traffic volumes, obtained through a simulation approach, were used as input parameters for a noise emission assessment conducted using the CNOSSOS-EU method, and an overall noise emissions reduction on the entire road network was found, even if its extent varied between road types

A sequence to sequence learning based car-following model for multi-step predictions considering reaction delay

Abstract: Car-following behavior modeling is of great importance for traffic simulation and analysis. Considering the multi-steps decision-making process in human driving, we propose a sequence to sequence (seq2seq) learning based car-following model incorporating not only memory effect but also reaction delay. Since the seq2seq architecture has the advantage of handling variable lengths of input and output sequences, in this paper, it is applied to car-following behavior modeling to memorize historical information and make multi-step predictions. We further compare the seq2seq model with a classical car-following model (IDM) and a deep learning car-following model (LSTM). The evaluation results indicate that the proposed model outperforms others for reproducing trajectory and capturing heterogeneous driving behaviors. Moreover, the platoon simulation demonstrates that the proposed model can well reproduce different levels of hysteresis phenomenon. The proposed model is further extended with spatial anticipation, which improves platoon simulation accuracy and traffic flow stability.

Potential last-mile impacts of crowdshipping services: a simulation-based evaluation

Abstract: Crowdsourced delivery services (crowdshipping) represent a shipping alternative to traditional delivery systems, particularly suitable for e-commerce. Although some benefits in terms of reduced pollution and congestion could be obtained by replacing dedicated freight trips, the impacts of crowdshipping are unclear and depend on several factors such as the transport mode used, the match between supply and demand, length of detours, and possible induced demand. For example, private drivers could modify their existing routes or engage in new trips to pick up and drop off packages; similarly, public transport users could carry along packages on their trips and drop them off at lockers installed around the stations. In this paper, we analyze by means of a simulation-based approach the potential impacts of alternative implementation frameworks. In order to account more realistically for last-mile delivery operations, a hybrid dynamic traffic simulation is adopted such that the macroscopic features of traffic (triggering of congestion, queue spillbacks and interactions with traffic signals) are reproduced in combination with the microscopic features of delivery operations (delivery vehicles are tracked along their routes). The effects on traffic and emissions are investigated for the adoption of crowdshipping by carriers delivering parcels in the city center of Rome, Italy. Results show that not only is the mode employed by crowdshippers crucial for the sustainability of such a measure, but also operational aspects involving the length of detour, parking behavior, and daily traffic variations. Crowdsourced deliveries by car have generally higher negative impacts than corresponding deliveries by public transit. However, limiting the deviations of crowdshippers from the original trips, providing adequate parking options, and incentivizing off-peak deliveries, could significantly reduce crowdshipping externalities.

Hybrid Spatio-Temporal Graph Convolutional Network: Improving Traffic Prediction with Navigation Data

Abstract: Traffic forecasting has recently attracted increasing interest due to the popularity of online navigation services, ridesharing and smart city projects. Owing to the non-stationary nature of road traffic, forecasting accuracy is fundamentally limited by the lack of contextual information. To address this issue, we propose the Hybrid Spatio-Temporal Graph Convolutional Network (H-STGCN), which is able to "deduce" future travel time by exploiting the data of upcoming traffic volume. Specifically, we propose an algorithm to acquire the upcoming traffic volume from an online navigation engine. Taking advantage of the piecewise-linear flow-density relationship, a novel transformer structure converts the upcoming volume into its equivalent in travel time. We combine this signal with the commonly-utilized travel-time signal, and then apply graph convolution to capture the spatial dependency. Particularly, we construct a compound adjacency matrix which reflects the innate traffic proximity. We conduct extensive experiments on real-world datasets. The results show that H-STGCN remarkably outperforms state-of-the-art methods in various metrics, especially for the prediction of non-recurring congestion.

Augmenting Vehicle Localization by Cooperative Sensing of the Driving Environment: Insight on Data Association in Urban Traffic Scenarios

Abstract: Precise vehicle positioning is a key element for the development of Cooperative Intelligent Transport Systems (C-ITS). In this context, we present a distributed processing technique to augment the performance of conventional Global Navigation Satellite Systems (GNSS) exploiting Vehicle-to-anything (V2X) communication systems. We propose a method, referred to as Implicit Cooperative Positioning with Data Association (ICP-DA), where the connected vehicles detect a set of passive features in the driving environment, solve the association task by pairing them with on-board sensor measurements and cooperatively localize the features to enhance the GNSS accuracy. We adopt a belief propagation algorithm to distribute the processing over the network, and solve both the data association and localization problems locally at vehicles. Numerical results on realistic traffic networks show that the ICP-DA method is able to significantly outperform the conventional GNSS. In particular, the analysis on a real urban road infrastructure highlights the robustness of the proposed method in real-life cases where the interactions among vehicles evolve over space and time according to traffic regulation mechanisms. Performances are investigated both in conventional traffic-light regulated scenarios and self-regulated environments (as representative of future automated driving scenarios) where vehicles autonomously cross the intersections taking gap-availability decisions for avoiding collisions. The analysis shows how the mutual coordination in platoons of vehicles eases the cooperation process and increases the positioning performance.

Development and Evaluation of an Adaptive Traffic Signal Control Scheme Under a Mixed-Automated Traffic Scenario

Abstract: This paper presents a novel adaptive traffic signal control scheme that addresses a mixed manual-automated traffic scenario in a typically isolated intersection. The traffic signals are optimized in a receding horizon control framework that aims at minimizing the total crossing time of all vehicles, considering their dynamical states. The control scheme ensures comfortable crossing of manually driven vehicles by retaining the basic features of the traditional signal management systems. The optimal signal changing times are broadcasted one cycle ahead, which enables the automated vehicles to tune their speed in order to cross the intersection with minimum stop-delay. More specifically, the framework optimizes the green time of each signal without considering the existing cycle-split concept explicitly. The proposed signal control scheme is evaluated in microscopic traffic simulation considering the different proportion of turning traffic at the intersection and various penetration rates of the automated vehicles. It is observed that the optimization process usually results in the shortest possible green period of each signal that can be realized without reducing the capacity of the intersection at any traffic volumes. Consequently, the resulting short signal cycle which is adaptive to the traffic around the intersection improves the average speed and reduces both the traffic density and the number of idling vehicles. As a consequence, the fuel consumption efficiency and the rate of CO2 emission around the intersection are also reduced. These results are compared with both the traditional fixed time and the actuated signal control schemes. As the portion of the automated vehicles increases in the case of the proposed scheme, the overall traffic flow performance further improves.

Game-Theoretic Modeling of Traffic in Unsignalized Intersection Network for Autonomous Vehicle Control Verification and Validation

Abstract: For a foreseeable future, autonomous vehicles (AVs) will operate in traffic together with human-driven vehicles. Their planning and control systems need extensive testing, including early-stage testing in simulations where the interactions among autonomous/human-driven vehicles are represented. Motivated by the need for such simulation tools, we propose a game-theoretic approach to modeling vehicle interactions, in particular, for urban traffic environments with unsignalized intersections. We develop traffic models with heterogeneous (in terms of their driving styles) and interactive vehicles based on our proposed approach, and use them for virtual testing, evaluation, and calibration of AV control systems. For illustration, we consider two AV control approaches, analyze their characteristics and performance based on the simulation results with our developed traffic models, and optimize the parameters of one of them.

A generic multi-level framework for microscopic traffic simulation with automated vehicles in mixed traffic

Abstract: With an increasing number of automated vehicles (AV) appearing on roads and interacting with conventional traffic, there is a need for improved simulation approaches to replicate and forecast the resulting effects. Interactions between AVs and their drivers, and interaction with other human drivers involve new types of complex behavioural processes. There is an increasing necessity to explicitly incorporate these human factor processes in simulation, which cannot be properly accounted for with most current models. In this paper, we present an extended conceptual simulation framework based on human factors processes and applicable for automated driving that does this. The framework makes use of previously constructed constructs to include the effects of driver task demand, situation awareness and fundamental diagrams of task demand to extend to automated driving. This is especially considered for the case of transition of control (ToC), as an important aspect of vehicle-driver interaction. The framework is demonstrated in two experimental cases that consider different ToC situations and is found to be face valid within the applied assumptions. Challenges remain in regard to a lack of quantitative evidence from traffic psychology, automated vehicle dynamics & control and human-vehicle interaction. With increasing amounts of research on-going in these areas, the extended framework will act as a valuable approach to further study and quantify the effects of AVs in mixed traffic in the future.

Simulation Strategies for Mixed Traffic Conditions: A Review of Car-Following Models and Simulation Frameworks

Abstract: The area of traffic flow modelling and analysis that bridges civil engineering, computer science, and mathematics has gained significant momentum in the urban areas due to increasing vehicular population causing traffic congestion and accidents. Notably, the existence of mixed traffic conditions has been proven to be a significant contributor to road accidents and congestion. The interaction of vehicles takes place in both lateral and longitudinal directions, giving rise to a two-dimensional (2D) traffic behaviour. This behaviour contradicts with the traditional car-following (CF) or one-dimensional (1D) lane-based traffic flow. Existing one-dimensional CF models did the inclusion of lane changing and overtaking behaviour of the mixed traffic stream with specific alterations. However, these parameters cannot describe the continuous lateral manoeuvre of mixed traffic flow. This review focuses on all the significant contributions made by 2D models in evaluating the lateral and longitudinal vehicle behaviour simultaneously. The accommodation of vehicle heterogeneity into the car-following models (homogeneous traffic models) is discussed in detail, along with their shortcomings and research gaps. Also, the review of commercially existing microscopic traffic simulation frameworks built to evaluate real-world traffic scenario are presented. This review identified various vehicle parameters adopted by existing CF models and whether the current 2D traffic models developed from CF models effectively captured the vehicle behaviour in mixed traffic conditions. Findings of this study are outlined at the end.

A generic simulation platform for cooperative adaptive cruise control under partially connected and automated environment

Abstract: Cooperative Adaptive Cruise Control (CACC) has been regarded as the most promising technology that can be carried out in field earliest under partially connected and automated environment, and is one of the few Connected and Automated Vehicle (CAV) applications closest to its final shape. Therefore, validation of CACC running in real traffic is in urgent need. To support the validation, simulation plays a key role. This paper proposes a next-generation simulation platform for CACC evaluation and validation. It is able to not only validate the performance of a CACC platoon by itself, but also evaluate its performance when being released into the traffic. In addition, it can also quantify the impact of CACC vehicles on transportation systems. The proposed platform is generic in terms of CACC control system, background traffic condition, road geometry, and traffic control scheme. Furthermore, it is capable of capturing the dynamics of platoon formation and disengagement when CACC equipped vehicles are mixed up with human-driven vehicles. The proposed platform is validated by comparing against actual field data and previously published theoretical studies. The results confirm the credibility of the proposed platform in terms of: i) CACC control system embedded; ii) CAV and Human-driven Vehicle (HV) mixed traffic simulation; iii) managed lane simulation. It is also revealed that current CACC technology without upgrade may be a safety hazard and permitting CACC only on a dedicated CACC lane might be a good first step for CACC commercialization at this moment.

Modeling and Prediction of Human Driver Behavior: A Survey.

Abstract: We present a review and taxonomy of 200 models from the literature on driver behavior modeling. We begin by introducing a mathematical formulation based on the partially observable stochastic game, which serves as a common framework for comparing and contrasting different driver models. Our taxonomy is constructed around the core modeling tasks of state estimation, intention estimation, trait estimation, and motion prediction, and also discusses the auxiliary tasks of risk estimation, anomaly detection, behavior imitation and microscopic traffic simulation. Existing driver models are categorized based on the specific tasks they address and key attributes of their approach.

Estimating multi-class dynamic origin-destination demand through a forward-backward algorithm on computational graphs

Abstract: Transportation networks are unprecedentedly complex with heterogeneous vehicular flow. Conventionally, vehicles are classified by size, the number of axles or engine types, e.g., standard passenger cars versus trucks. However, vehicle flow heterogeneity stems from many other aspects in general, e.g., ride-sourcing vehicles versus personal vehicles, human driven vehicles versus connected and automated vehicles. Provided with some observations of vehicular flow for each class in a large-scale transportation network, how to estimate the multi-class spatio-temporal vehicular flow, in terms of time-varying Origin-Destination (OD) demand and path/link flow, remains a big challenge. This paper presents a solution framework for multi-class dynamic OD demand estimation (MCDODE) in large-scale networks that work for any vehicular data in general. The proposed framework cast the standard OD estimation methods into a computational graph with tensor representations of spatio-temporal flow and all intermediate features involved in the MCDODE formulation. A forward-backward algorithm is proposed to efficiently solve the MCDODE formulation on computational graphs. In addition, we propose a novel concept of tree-based cumulative curves to compute the exact multi-class Dynamic Assignment Ratio (DAR) matrix. A Growing Tree algorithm is developed to construct tree-based cumulative curves. The proposed framework is examined on a small network, a mid-size network as well as a real-world large-scale network. The experiment results indicate that the proposed framework is compelling, satisfactory and computationally plausible.

A data-driven approach to calibrate microsimulation models based on the degree of saturation at signalized intersections

Abstract: Microscopic traffic simulation is considered as a reliable tool in transportation planning and management. Rational solutions from such simulations are contingent upon how well the simulation software is calibrated and validated to replicate real-world road network scenarios. Most of the existing calibration and validation efforts are normally based on the comparative analysis between the built-in attributes of VISSIM and the real-world scenarios using measures of effectiveness (MOEs). VISSIM attributes such as the volume-to-capacity ratios, vehicle delay, and queue lengths, are normally used during the validation process of signalized intersections. However, validating VISSIM based on a non-inbuilt attribute is yet to be explored. This paper proposes a step-by-step procedure for calibrating signalized intersections in VISSIM based on a measurable variable, which is the degree of saturation. The approach was applied to a case study of four signalized intersections in Miami, Florida. The methodology utilized real-world vehicle trajectory data to determine the optimal values of VISSIM car-following parameters required for calibration. Statistical results revealed that both the saturation headways obtained from VISSIM and the saturation headways observed in the field follow the same distribution. The results signify that VISSIM could be calibrated using a non-inbuilt attribute, and moreover generates accurate data compared to the field measurements.

Mixed reality test environment for autonomous cars using Unity 3D and SUMO

Abstract: Developing a vehicle is always a long and complex task. This is especially true for autonomous cars. Tasks performed by the driver are taken over by the vehicle and must be performed with maximum reliability. Developing these systems is a difficult task, especially due to limited testing capabilities. Testing a vehicle in a closed environment is safe and controlled, but the variety of test scenarios is limited. By using mixed reality environments, one can create a diverse environment around a real test vehicle, with traffic, obstacles, and unexpected situations. The real movement of the test vehicle allows testing decision and motion planning level vehicular functions. Information from the virtual world can be considered as input to the vehicle sensor. Mixed reality or digital twin simulation environments can greatly assist the autonomous vehicle development process and also serve as a basis for validation procedures for such systems. This article introduces a mixed reality simulation environment that integrates a real test vehicle into a virtual environment, can handle other real and virtual obstacles, contains traffic simulations, and visualizes all of these.

Adaptive Control Strategies for Urban Network Traffic via a Decentralized Approach With User-Optimal Routing

Abstract: This paper presents an adaptive linear quadratic optimal traffic control system. The control strategies are solved via a decentralized approach and complemented with a user-optimal network traffic router. The user-optimal routing algorithm assists drivers respond to prevailing traffic state and control settings and seek the quickest route toward their destinations. The proposed control system is implemented and tested over different scenario settings including a real life scenario in Central London, UK. The study reveals that the proposed system could coverage to a performance similar to its centralized counterpart with the routing algorithm even under congested conditions. This highlights the potential of decentralized control with effective travel guidance in cooperative traffic management.

A Taxonomy and Review of Algorithms for Modeling and Predicting Human Driver Behavior

Abstract: We present a review and taxonomy of 200 models from the literature on driver behavior modeling. We begin by introducing a mathematical framework for describing the dynamics of interactive multi-agent traffic. Based on the partially observable stochastic game, this framework provides a basis for discussing different driver modeling techniques. Our taxonomy is constructed around the core modeling tasks of state estimation, intention estimation, trait estimation, and motion prediction, and also discusses the auxiliary tasks of risk estimation, anomaly detection, behavior imitation and microscopic traffic simulation. Existing driver models are categorized based on the specific tasks they address and key attributes of their approach.

A two-level probabilistic approach for validation of stochastic traffic simulations: impact of drivers' heterogeneity models

Abstract: This paper shows how traffic heterogeneity, and the way it is encoded into a model, drastically affects a model ability to reproduce observed traffic. Being heterogeneity a major source of uncertainty, to correctly frame the proposed validation methodology we have first reviewed and adapted cross-disciplinary theoretical concepts about uncertainty modelling to traffic simulation. A number of open issues, including error compensation and model overfitting, has been interpreted and clarified through the proposed framework. A two-level probabilistic approach has been applied to run stochastic simulations of three NGSIM I-80 traffic scenarios, and quantitatively infer the impact of heterogeneity. According to this approach, both the car-following and the lane-changing models of each vehicle have been calibrated against observed trajectories. Based on the estimated parameters distributions, different models of heterogeneity have been quantitatively validated against macroscopic traffic patterns. Being traffic a collective phenomenon emerging from microscopic interactions, even models calibrated on microscopic trajectories need to be quantitatively validated on macroscopic traffic patterns too. Among other results, normal distributions of the model parameters, which are customarily applied in traffic simulation practice, have been found unable to reproduce the observed congestion patterns. Parameters correlation, being claimed as highly influential in previous works, is responsible for a model overfitting in traffic scenarios with low congestion. In the end, it has been demonstrated that a thorough characterization of parameters heterogeneity cannot be left out in traffic simulation, if an ersatz representation of traffic is to be avoided.

Assessing the impacts of shared autonomous vehicles on congestion and curb use: A traffic simulation study in The Hague, Netherlands

Abstract: New developments in the automotive world have the power to change mobility, but because of high uncertainties, municipalities are adopting a wait-and-see attitude. Nonetheless, autonomous, connected and shared vehicle technologies are in a far stage of development and it is only a matter of time before shared autonomous vehicles (SAVs) enter urban traffic. This research aims to provide insights into the congestion effects of SAVs on urban traffic, focusing on the differences in microscopic behaviour from conventional cars, and to investigate which easy-to-implement solutions a municipality could apply to facilitate the new mix of traffic. This was researched by performing a simulation study, using the traffic simulation package Vissim and a case study of a network in the city of The Hague during the morning peak in 2040. Several SAV market penetration scenarios were tested: 0%, 3%, 25%, 50% and 100% SAV usage by travellers. Additionally, two network designs were implemented: dedicated lanes for SAVs and kiss & ride (K&R)-facilities. From the results, it was clear that while the autonomous driving capabilities of SAVs help reduce traffic congestion, they also have a negative effect by stopping on the curbside to drop off passengers, forming bottlenecks for other road users, and by circulating on the network using low capacity links. Below the line, this adds up to an overall negative effect on urban traffic congestion according to our results. The dedicated lanes design was unsuccessful at reducing this congestion caused by SAVs. The K&R design, however, was successful at reducing delays, but only for SAV penetration rates higher than 25%. These exact effects are not generalizable due to limitations in network size and simulation software. However, the results can be seen as indicative for planning purposes. Similar effects could be expected in cities where transport network companies (TNCs) such as Uber become exceptionally popular with non-autonomous cars. The advice for municipalities is to closely monitor the situation and to account for SAVs (and TNCs) in each new infrastructural project.

Mixed Test Environment-based Vehicle-in-the-loop Validation - A New Testing Approach for Autonomous Vehicles

Abstract: The current test of autonomous driving technology requires extensive experimental verification, whether in simulation or on real roads. Netherthless, how to test autonomous vehicles thoroughly in a safe and comprehensive manner remains a major challenge. To achieve safer and more effective autonomous driving testing, this paper proposes a novel mixed test environment-based validation method with vehicle in the loop(ViL) : (1) our method supports more realistic drive safety tests in mixed scenarios which integrate the synthetic and the real-world scenarios. Synthetic scenarios offer complex traffic simulation with diverse road conditions. The real-world scenarios introduce the real autonomous driving vehicle, the real sensor suite as well as the test field to the test loop, having further bridged the gap between Hardware-in-the-Loop(HiL) testing and real road tests than ViL; (2) virtual perceptional results are simulated directly and delivered to the real vehicle in Unified Fusion Data Format(UFDF), without rendering virtual detection data for reduced resource consumption; (3) diverse test scenarios are configurable and reproducible with OSM-based High Definition(HD) map, enabling the simulation to be decoupled from a specific test filed or traffic facilities. A series of experiments on the application of our method have been demonstrated, and our approach is proved to be a promising drive safety testing technique before actual road testing.

Multi-objective eco-routing for dynamic control of connected & automated vehicles

Abstract: The advent of intelligent vehicles that can communicate with infrastructure as well as automate the movement provides a range of new options to address key urban traffic issues such as congestion and pollution, without the need for centralized traffic control. Furthermore, the advances in the information, communication, and sensing technologies have provided access to real-time traffic and emission data. Leveraging these advancements, a dynamic multi-objective eco-routing strategy for connected & automated vehicles (CAVs) is proposed and implemented in a distributed traffic management system. It is applied to the road network of downtown Toronto in an in-house agent-based traffic simulation platform. The performance of the proposed system is compared to various single-objective optimizations. Simulation results show the significance of incorporating real-time emission and traffic state into the dynamic routing, along with considering the expected delays at the downstream intersections. The proposed multi-objective eco-routing has the potential of reducing GHG and NOx emissions by 43% and 18.58%, respectively, while reducing average travel time by 40%.

Game-Theoretic Modeling of Multi-Vehicle Interactions at Uncontrolled Intersections

Abstract: Motivated by the need for simulation tools for testing, verification and validation of autonomous driving systems that operate in traffic consisting of both autonomous and human-driven vehicles, we propose a game-theoretic framework for modeling the interactive behavior of vehicles at uncontrolled intersections. The proposed vehicle interaction model is based on a novel formulation of dynamic games with multiple concurrent leader-follower pairs, induced from common traffic rules. Based on simulation results for various intersection scenarios, we show that the model exhibits reasonable behavior expected in traffic, including the capability of reproducing scenarios extracted from real-world traffic data and reasonable performance in resolving traffic conflicts. The model is further validated based on the level-of-service traffic quality rating system and demonstrates manageable computational complexity compared to traditional multi-player game-theoretic models.

A Comparison of Deep Learning Methods for Urban Traffic Forecasting using Floating Car Data

Abstract: Cities today must address the challenge of sustainable mobility, and traffic state forecasting plays a key role in mitigating traffic congestion in urban areas. For example, predicting path travel time is a crucial issue in navigation and route planning applications. Furthermore, the pervasive penetration of information and communication technologies makes floating car data an important source of real-time data for intelligent transportation system applications. This paper deals with the problem of forecasting urban traffic when floating car data is available. A comparison of four deep learning methods is presented to demonstrate the capabilities of the neural network approaches (recurrent and/or convolutional) in solving the traffic forecasting problem in an urban context. Different tests are proposed in order to not only evaluate the developed deep learning models, but also to analyze how the penetration rates of floating cars affect forecasting accuracy. The presented experiments were designed according to a microscopic traffic simulation approach in order to emulate floating car data fleets, which provide vehicle position and speed, and to validate the obtained results. Finally, some conclusions and further research are presented.

Short‐term prediction of traffic flow under incident conditions using graph convolutional recurrent neural network and traffic simulation

Abstract: The objective of the study is to predict traffic flow under unusual conditions by using a deep learning model. Conventionally, machine-learning-based traffic prediction is frequently carried out. Model learning requires large amounts of training data; however, collecting sufficient samples is a challenge in the event of traffic incidents. To address this challenge, large amounts of traffic data were generated by performing traffic simulations under various traffic incidents. These data were used as training data, and a deep learning model with graph convolution and input of traffic incident information features was proposed. Subsequently, the prediction accuracy was compared with other models such as long short-term memory, which is typically used in traffic prediction. The results demonstrated the superiority of the proposed model in representing phenomena with strong spatio-temporal dependencies, such as traffic flow, and its effectiveness in traffic prediction.

Detecting, Tracking and Counting People Getting On/Off a Metropolitan Train Using a Standard Video Camera.

Abstract: The main source of delays in public transport systems (buses, trams, metros, railways) takes place in their stations. For example, a public transport vehicle can travel at 60 km per hour between stations, but its commercial speed (average en-route speed, including any intermediate delay) does not reach more than half of that value. Therefore, the problem that public transport operators must solve is how to reduce the delay in stations. From the perspective of transport engineering, there are several ways to approach this issue, from the design of infrastructure and vehicles to passenger traffic management. The tools normally available to traffic engineers are analytical models, microscopic traffic simulation, and, ultimately, real-scale laboratory experiments. In any case, the data that are required are number of passengers that get on and off from the vehicles, as well as the number of passengers waiting on platforms. Traditionally, such data has been collected manually by field counts or through videos that are then processed by hand. On the other hand, public transport networks, specially metropolitan railways, have an extensive monitoring infrastructure based on standard video cameras. Traditionally, these are observed manually or with very basic signal processing support, so there is significant scope for improving data capture and for automating the analysis of site usage, safety, and surveillance. This article shows a way of collecting and analyzing the data needed to feed both traffic models and analyze laboratory experimentation, exploiting recent intelligent sensing approaches. The paper presents a new public video dataset gathered using real-scale laboratory recordings. Part of this dataset has been annotated by hand, marking up head locations to provide a ground-truth on which to train and evaluate deep learning detection and tracking algorithms. Tracking outputs are then used to count people getting on and off, achieving a mean accuracy of 92% with less than 0.15% standard deviation on 322 mostly unseen dataset video sequences.

Multi-agent deep neural networks coupled with LQF-MWM algorithm for traffic control and emergency vehicles guidance

Abstract: Authorities in modern cities are facing daily challenges related to traffic control. Due to the problem complexity caused by the urbanization growth, investing in developing traffic signal control systems (TSCS) to guarantee better mobility has taken more attention by these authorities. In the existing literature, the majority of TSCS offers only a real-time control for a detected traffic problem without considering early prediction and estimation of its occurrence. Furthermore, traffic problems related to the arrival and guidance of emergency vehicles are rarely considered. Based on these gaps, we rely on concepts and mechanisms from both, the Artificial and the convolution neural networks (ANN and CNN), coupled with the longest queue first maximal weight matching algorithm (LQF-MWM), to develop PANNAL, a predictive and reactive TSCS. PANNAL is a Multi-Agent based System, where each individual agent has ANN, CNN, and LQF-MWM to adapt signal sequences and durations and favor the crossing of emergency vehicles. Agents have a heterarchical architecture considered for coordination. We leant on VISSIM, a state-of-the-art traffic simulation software for simulation and evaluation. We adopted algorithms, scenarios, key performance indicators, and evaluation results from the recent literature for benchmarking. These algorithms are pre-emptive and have a high performance and competitive results in traffic control of disturbed traffic condition.