**Are you an EPFL student looking for a semester project?**

Work with us on data science and visualisation projects, and deploy your project as an app on top of GraphSearch.

MOOC# Intro to Traffic Flow Modeling and Intelligent Transport Systems

Description

Learn how to describe, model and control urban traffic congestion in simple ways and gain insight into advanced traffic management schemes that improve mobility in cities and highways.

Official source

This page is automatically generated and may contain information that is not correct, complete, up-to-date, or relevant to your search query. The same applies to every other page on this website. Please make sure to verify the information with EPFL's official sources.

Instructors (2)

Related courses (37)

Related publications (40)

Related concepts (139)

CIVIL-349: Traffic engineering

Introduce the major elements of transportation systems and traffic engineering-Develop analytical and technical skills in applying the fundamentals of the transport field-Understand the key concepts a

ME-427: Networked control systems

This course offers an introduction to control systems using communication networks for interfacing sensors, actuators, controllers, and processes. Challenges due to network non-idealities and opportun

CS-411: Digital education

This course addresses the relationship between specific technological features and the learners' cognitive processes. It also covers the methods and results of empirical studies on this topic: do stud

Transport network analysis

A transport network, or transportation network, is a network or graph in geographic space, describing an infrastructure that permits and constrains movement or flow. Examples include but are not limited to road networks, railways, air routes, pipelines, aqueducts, and power lines. The digital representation of these networks, and the methods for their analysis, is a core part of spatial analysis, geographic information systems, public utilities, and transport engineering.

Urban area

An urban area, built-up area or urban agglomeration is a human settlement with a high population-density and an infrastructure of built environment. This is the core of a metropolitan statistical area in the United States, if it contains a population of more than 50,000. Urban areas originate through urbanization, and researchers categorize them as cities, towns, conurbations or suburbs. In urbanism, the term "urban area" contrasts to rural areas such as villages and hamlets; in urban sociology or urban anthropology it contrasts with natural environment.

Fundamental diagram of traffic flow

The fundamental diagram of traffic flow is a diagram that gives a relation between road traffic flux (vehicles/hour) and the traffic density (vehicles/km). A macroscopic traffic model involving traffic flux, traffic density and velocity forms the basis of the fundamental diagram. It can be used to predict the capability of a road system, or its behaviour when applying inflow regulation or speed limits. There is a connection between traffic density and vehicle velocity: The more vehicles are on a road, the slower their velocity will be.

Modeling urban traffic on the level of network is a wide research area oriented to the development of ITS. In this thesis properties of models based on MFD (Macroscopic Fundamental Diagram) are studied. The idea behind MFD is to say that the state of the traffic inside an urban zone is fully determined by its accumulation (the number of traveling vehicles) and that the dynamics of accumulation is caused by the inflow of vehicles (flow of vehicles that enter the zone or start their trips from inside). Nowadays, two different philosophies of modeling dynamics of accumulation exist in the literature. The first one (outflow-MFD) postulates that the outflow of vehicles depends on accumulation. The second one (speed-MFD) postulates that the space-mean speed of vehicles depends on accumulation. The second philosophy already has strong empirical support based on observations of traffic inside many big (kilometer-scale) urban areas around the world. Thus, different speed-MFD models are of great scientific interest.
The thesis is mainly devoted to the comparison of so-called PL model (which assumes the existence of both speed-MFD and outflow-MFD) and TB model (which assumes the equality of speeds of vehicles and explicitly assumes the existence of trip length distribution). It was shown that PL model cannot accurately describe the dynamics of accumulation after the jump of inflow. In this case TB model is more preferable. Moreover, it was proven that PL model is a specific case of TB model for the exponential trip length distribution. This makes TB model more attractive than PL model for the practical usage.
TB model can be formulated mathematically either as integral equation or nonlocal PDE. Thus, the main drawback of TB model that can be an obstacle in practice is its computational complexity. In this thesis it was proposed to approximate TB model with a simpler model which does not require precise information about the trip length distribution. This so-called M model operates with only the mean and the standard deviation of distribution and has a form of ODE. The analytical comparison between PL, TB and M models proved that M model is much closer to TB model than PL model in the case of constant speed-MFD. The more realistic case of decreasing speed-MFD was studied through the numerical tests and also showed the same effect. Thus, the main conclusion of the study is that M model has practical potential as an elegant and computationally cheap approximation of TB model. Also, given that in the case of constant speed TB model is a type of LTI system, it can be expected that M model might be useful for a wide range of problems that are not related to transportation. However, in this thesis the conclusion about the small difference between M and TB models was made for the inflows that are typical for the transportation field. More precisely, only peak hour shaped (smooth and with small jumps) functions were studied.
Despite the simple form of M model it was found that there exists even more simple ODE approximation of TB model. This so-called $\alpha$ model works quite well for both smooth and jumping inflows except the case of a short period of time following the jump of inflow. Thus, it might be another good alternative to TB model. The main advantage of $\alpha$ model in the case of realistic speed-MFD function is its convex formulation which allows $\alpha$ model to be efficiently used inside optimization frameworks.

Nikolaos Geroliminis, Dimitrios Tsitsokas, Anastasios Kouvelas

Dedicated bus lanes provide a low cost and easily implementable strategy to improve transit service by minimizing congestion-related delays. Identifying the best spatial distribution of bus-only lanes in order to maximize traffic performance of an urban network while balancing the trade-off between bus priority and regular traffic disturbance is a challenging task. This paper studies the problem of optimal dedicated bus lane allocation and proposes a modeling framework based on a link-level dynamic traffic modeling paradigm, which is compatible with the dynamic characteristics of congestion propagation that can be correlated with bus lane relative positions. The problem is formulated as a non-linear combinatorial optimization problem with binary variables. An algorithmic scheme based on a problem-specific heuristic and Large Neighborhood Search metaheuristic, potentially combined with a network decomposition technique and a performance-based learning process for increased efficiency, is proposed for deriving good quality solutions for large-scale network instances. Numerical application results for a real city center demonstrate the efficiency of the proposed framework in finding effective bus lane network configurations; when compared to the initial network state they exhibit the potential of bus lanes to improve travel time for car and bus users.

2021Lectures in this MOOC (25)

This dissertation introduces traffic forecasting methods for different network configurations and data availability.Chapter 2 focuses on single freeway cases.Although its topology is simple, the non-linearity of traffic features makes this prediction still a challenging task.We propose the dynamic linear model (DLM) to approximate the non-linear traffic features. Unlike a static linear regression model, the DLM assumes that its parameters change over time.We design the DLM with time-dependent model parameters to describe the spatiotemporal characteristics of time-series traffic data. Based on our DLM and its model parameters analytically trained using historical data, we suggest the optimal linear predictor in the minimum mean square error (MMSE) sense.We compare our prediction accuracy by estimating expected travel time based on the traffic prediction for freeways in California (I210-E and I5-S) under highly congested traffic conditions with other baselines. We show significant improvements in accuracy, especially for short-term prediction.Chapter 3 aims to generalize the DLM to extensive freeway networks with more complex topologies.Most resources would be consumed to estimate unnecessary spatiotemporal correlations if the DLM was directly used for a large-scale network.Defining features on graphs relaxes such issues by cutting unnecessary connections in advance based on predefined topology information.Exploiting the graph signal processing, we represent traffic dynamics over freeway networks using multiple graph heat diffusion kernels and integrate the kernels into DLM with Bayes' rule. We optimize the model parameters using Bayesian inference to minimize the prediction errors.The proposed model demonstrates prediction accuracy comparable to state-of-the-art deep neural networks with lower computational effort. It notably achieves excellent performance for long-term prediction through the inheritance of periodicity modeling in DLM.Chapter 4 proposes a deep neural network model to predict traffic features on large-scale freeway networks.These days, deep learning methods have heavily tackled traffic forecasting problems of freeway networks because they are outstanding at learning highly complex correlations between variables both in time and space, which the linear models might be limited to.Adopting a graph convolutional network (GCN) becomes a standard to extract spatial correlations; therefore, most works have achieved great prediction accuracy by implanting it into their architecture.However, the conventional GCN has the drawback that receptive field size should be small, i.e., barely refers to traffic features of remote sensors, resulting in inaccurate long-term prediction.We suggest a forecasting model called two-level resolution deep neural network (TwoResNet) that overcomes the limitation.It consists of two resolution blocks:The low-resolution block predicts traffic on a macroscopic scale, such as regional traffic changes.On the other hand, the high-resolution block predicts traffic on a microscopic scale by using GCN to extract spatial correlations, referring to the regional changes produced by the low-resolution block.This process allows the GCN to refer to the traffic features from remote sensors.As a result, TwoResNet achieves competitive prediction accuracy compared to state-of-the-art methods, especially showing excellent performance for long-term predictions.

Traffic flow variablesMOOC: Intro to Traffic Flow Modeling and Intelligent Transport Systems

Covers traffic flow variables, including space-time diagrams and speed calculations, based on Edie (1963).

Fundamental Diagram: Traffic Flow Modeling and ITSMOOC: Intro to Traffic Flow Modeling and Intelligent Transport Systems

Covers the fundamental diagram in traffic flow modeling and discusses field observations, empirical data, speed-density relations, bottlenecks, and capacity drop phenomena.

Input-Output DiagramsMOOC: Intro to Traffic Flow Modeling and Intelligent Transport Systems

Explains input-output diagrams in queueing systems and traffic signals, including Little's Formula and issues like FIFO violation and noise effects.

Traffic flow modeling: Developing and testing traffic modelsMOOC: Intro to Traffic Flow Modeling and Intelligent Transport Systems

Covers the fundamentals of traffic flow modeling, including model classification and testing techniques.

Traffic Flow Modeling: An Introduction to LWR TheoryMOOC: Intro to Traffic Flow Modeling and Intelligent Transport Systems

Covers the LWR traffic flow theory, including conservation laws, shock waves, basics, and principles.