Vadim Sokolov

Deep learning for short-term traffic flow prediction

Abstract We develop a deep learning model to predict traffic flows. The main contribution is development of an architecture that combines a linear model that is fitted using l 1 regularization and a sequence of tanh layers. The challenge of predicting traffic flows are the sharp nonlinearities due to transitions between free flow, breakdown, recovery and congestion. We show that deep learning architectures can capture these nonlinear spatio-temporal effects. The first layer identifies spatio-temporal relations among predictors and other layers model nonlinear relations. We illustrate our methodology on road sensor data from Interstate I-55 and predict traffic flows during two special events; a Chicago Bears football game and an extreme snowstorm event. Both cases have sharp traffic flow regime changes, occurring very suddenly, and we show how deep learning provides precise short term traffic flow predictions.

Deep Learning: A Bayesian Perspective

Deep learning is a form of machine learning for nonlinear high dimensional pattern matching and prediction. By taking a Bayesian probabilistic perspective, we provide a number of insights into more efficient algorithms for optimisation and hyper-parameter tuning. Traditional high-dimensional data reduction techniques, such as principal component analysis (PCA), partial least squares (PLS), reduced rank regression (RRR), projection pursuit regression (PPR) are all shown to be shallow learners. Their deep learning counterparts exploit multiple deep layers of data reduction which provide predictive performance gains. Stochastic gradient descent (SGD) training optimisation and Dropout (DO) regularization provide estimation and variable selection. Bayesian regularization is central to finding weights and connections in networks to optimize the predictive bias-variance trade-off. To illustrate our methodology, we provide an analysis of international bookings on Airbnb. Finally, we conclude with directions for future research.

Analysis of the Effects of Connected–Automated Vehicle Technologies on Travel Demand

Connected–automated vehicle (CAV) technologies are likely to have significant effects not only on how vehicles operate in the transportation system, but also on how individuals behave and use their vehicles. While many CAV technologies—such as connected adaptive cruise control and ecosignals—have the potential to increase network throughput and efficiency, many of these same technologies have a secondary effect of reducing driver burden, which can drive changes in travel behavior. Such changes in travel behavior—in effect, lowering the cost of driving—have the potential to increase greatly the utilization of the transportation system with concurrent negative externalities, such as congestion, energy use, and emissions, working against the positive effects on the transportation system resulting from increased capacity. To date, few studies have analyzed the potential effects on CAV technologies from a systems perspective; studies often focus on gains and losses to an individual vehicle, at a single intersection, or along a corridor. However, travel demand and traffic flow constitute a complex, adaptive, nonlinear system. Therefore, in this study, an advanced transportation systems simulation model—POLARIS—was used. POLARIS includes cosimulation of travel behavior and traffic flow to study the potential effects of several CAV technologies at the regional level. Various technology penetration levels and changes in travel time sensitivity have been analyzed to determine a potential range of effects on vehicle miles traveled from various CAV technologies.

Bayesian analysis of traffic flow on interstate I-55: The LWR model

Transportation departments take actions to manage traffic flow and reduce travel times based on estimated current and projected traffic conditions. Travel time estimates and forecasts require information on traffic density which are combined with a model to project traffic flow such as the Lighthill-Whitham-Richards (LWR) model. We develop a particle filtering and learning algorithm to estimate the current traffic density state and the LWR parameters. These inputs are related to the so-called fundamental diagram, which describes the relationship between traffic flow and density. We build on existing methodology by allowing real-time updating of the posterior uncertainty for the critical density and capacity parameters. Our methodology is applied to traffic flow data from interstate highway I-55 in Chicago. We provide a real-time data analysis of how to learn the drop in capacity as a result of a major traffic accident. Our algorithm allows us to accurately assess the uncertainty of the current traffic state at shock waves, where the uncertainty is a mixture distribution. We show that Bayesian learning can correct the estimation bias that is present in the model with fixed parameters.

Bayesian Optimization for Transportation Simulators

Abstract Mobility dynamics in urban transportation systems is governed by a large number of travelers that act according to their utilities, preferences and biases. The mobility patterns that we observe are the results of the emerging traveler’s behaviors and, in practice, we develop models that represents mobility patterns and their resulting traffic flows. Models, however, can only approximate the processes they represent and often times do not reproduce exact matches to the true system’s observed data. Systematic adjustments, or calibrations, to the model and its input variables may be required to align the associated outputs more closely with their true values. In this paper we outline a mathematical framework that allows the calibration for parameters of urban transportation models through a distributed, Gaussian Process Bayesian regression with active learning methods and demonstrate using a ground transportation simulation model.

Maximization of Platoon Formation Through Centralized Routing and Departure Time Coordination

Platooning allows vehicles to travel with a small intervehicle distance in a coordinated fashion because of vehicle-to-vehicle connectivity. When applied at a larger scale, platooning creates significant opportunities for energy savings because of reduced aerodynamic drag, as well as increased road capacity and a reduction in congestion resulting from shorter vehicle headways. These potential savings are maximized, however, if platooning-capable vehicles spend most of their travel time within platoons. Ad hoc platoon formation may not ensure a high rate of platoon driving. This paper considers the problem of central coordination of platooning-capable vehicles. Coordination of their routes and departure times can maximize the fuel savings afforded by platooning vehicles. The resulting problem is a combinatorial optimization problem that considers the platoon coordination and vehicle routing problems simultaneously. The methodology is demonstrated through evaluation of the benefits of a coordinated solution and comparison with the uncoordinated case when platoons form only in an ad hoc manner. The coordinated and uncoordinated scenarios are compared on a grid network with various assumptions about demand and the time vehicles are willing to wait.

Assessing the energy impact of traffic management and vehicle hybridisation

We provide a review of methodologies previously used to evaluate impacts of transportation systems and changes in transportation infrastructure on energy consumption. We present a new framework that allows estimating the energy impacts of managed traffic lanes in the context of vehicle automation. The presented framework relies on two major components, an integrated transportation system simulator and a powertrain simulator. For the transportation system simulator we propose using integrated transportation system simulator POLARIS. For the powertrain simulator we use AUTONOMIE, a tool funded by the US Department of Energy. Both tools are developed at Argonne National Laboratory. We demonstrate our approach by modelling a transportation corridor along a major highway. Two scenarios are considered, unmanaged, when both trucks and cars use all the lanes of the highway and managed, under which one of the highway lanes is a dedicated lane for truck traffic and trucks are forming platoons using adaptive cruise control technology. We provide the numerical results of the experiment at the end of the paper. We also present the impact of vehicle hybridisation combined with automation on the energy consumption.