Data Mining Algorithms for Traffic Sampling, Estimation and Forecasting
Committee memberObradovic, Zoran
DepartmentComputer and Information Science
Permanent link to this recordhttp://hdl.handle.net/20.500.12613/2720
MetadataShow full item record
AbstractDespite the significant investments over the last few decades to enhance and improve road infrastructure worldwide, the capacity of road networks has not kept pace with the ever increasing growth in demand. As a result, congestion has become endemic to many highways and city streets. As an alternative to costly and sometimes infeasible construction of new roads, transportation departments are increasingly looking at ways to improve traffic flow over the existing infrastructure. The biggest challenge in accomplishing this goal is the ability to sample traffic data, estimate traffic current state, and forecast its future behavior. In this thesis, we first address the problem of traffic sampling where we propose strategies for frugal sensing where we collect a fraction of the observed traffic information to reduce costs while achieving high accuracy. Next we demonstrate how traffic estimation using deterministic traffic models can be improved using proposed data reconstruction techniques. Finally, we propose how mixture of experts algorithm which consists of two regime-specific linear predictors and a decision tree gating function can improve short-term and long-term traffic forecasting. As mobile devices become more pervasive, participatory sensing is becoming an attractive way of collecting large quantities of valuable location-based data. An important participatory sensing application is traffic monitoring, where GPS-enabled smartphones can provide invaluable information about traffic conditions. We propose a strategy for frugal sensing in which the participants send only a fraction of the observed traffic information to reduce costs while achieving high accuracy. The strategy is based on autonomous sensing, in which participants make decisions to send traffic information without guidance from the central server, thus reducing the communication overhead and improving privacy. To provide accurate and computationally efficient estimation of the current traffic, we propose to use a budgeted version of the Gaussian Process model on the server side. The experiments on real-life traffic data sets indicate that the proposed approach can use up to two orders of magnitude less samples than a baseline approach with only a negligible loss in accuracy. The estimation of the state of traffic provides a detailed picture of the conditions of a traffic network based on limited traffic measurements and, as such, plays a key role in intelligent transportation systems. Most often, traffic measurements are aggregated over multiple time steps, and this procedure raises the question of how to best use this information for state estimation. Reconstructing the high-resolution measurements from the aggregated ones and using them to correct the state estimates at every time step are proposed. Several reconstruction techniques from signal processing, including kernel regression and a reconstruction approach based on convex optimization, were considered. Experimental results show that signal reconstruction leads to more accurate traffic state estimation as compared with the standard approach for dealing with aggregated measurements. Accurate traffic speed forecasting can help in trip planning by allowing travelers to avoid congested routes, either by choosing alternative routes or by changing the departure time. An important feature of traffic is that it consists of free flow and congested regimes, which have significantly different properties. Training a single traffic speed predictor for both regimes typically results in suboptimal accuracy. To address this problem, a mixture of experts algorithm which consists of two regime-specific linear predictors and a decision tree gating function was developed. Experimental results showed that mixture of experts approach outperforms several popular benchmark approaches.
ADA complianceFor Americans with Disabilities Act (ADA) accommodation, including help with reading this content, please contact firstname.lastname@example.org
Showing items related by title, author, creator and subject.
Data-Fitted Generic Second Order Macroscopic Traffic Flow ModelsSeibold, Benjamin; Szyld, Daniel; Chidyagwai, Prince; Piccoli, Benedetto, 1968- (Temple University. Libraries, 2013)The Aw-Rascle-Zhang (ARZ) model has become a favorable ``second order" macroscopic traffic model, which corrects several shortcomings of the Payne-Whitham (PW) model. The ARZ model possesses a family of flow rate versus density (FD) curves, rather than a single one as in the ``first order" Lighthill-Whitham-Richards (LWR) model. This is more realistic especially during congested traffic state, where the historic fundamental diagram data points are observed to be set-valued. However, the ARZ model also possesses some obvious shortcomings, e.g., it assumes multiple maximum traffic densities which should be a ``property" of road. Instead, we propose a Generalized ARZ (GARZ) model under the generic framework of ``second order" macroscopic models to overcome the drawbacks of the ARZ model. A systematic approach is presented to design generic ``second order" models from historic data, e.g., we construct a family of flow rate curves by fitting with data. Based on the GARZ model, we then propose a phase-transition-like model that allows the flow rate curves to coincide in the free flow regime. The resulting model is called Collapsed GARZ (CGARZ) model. The CGARZ model keeps the flavor of phase transition models in the sense that it assume a single FD function in the free-flow phase. However, one should note that there is no real phase transition in the CGARZ model. To investigate to which extent the new generic ``second order" models (GARZ, CGARZ) improve the prediction accuracy of macroscopic models, we perform a comparison of the proposed models with two types of LWR models and their ``second order" generalizations, given by the ARZ model, via a three-detector problem test. In this test framework, the initial and boundary conditions are derived from real traffic data. In terms of using historic traffic data, a statistical technique, the so-called kernel density estimation, is applied to obtain density and velocity distributions from trajectory data, and a cubic interpolation is employed to formulate boundary condition from single-loop sensor data. Moreover, a relaxation term is added to the momentum equation of selected ``second order" models to address further unrealistic aspects of homogeneous models. Using these inhomogeneous ``second order" models, we study which choices of the relaxation term &tau are realistic.
The Effect of Historic Paving Materials on Traffic SpeedNogueira, Xavier; Mennis, Jeremy; 0000-0002-3458-1646; 0000-0001-6319-8622 (2019-10-01)Slowing traffic speed in urban areas has been shown to reduce pedestrian injuries and fatalities due to automobile accidents. This research aims to measure how brick and granite block paving materials, which were widely used historically prior to the use of asphalt paving in many cities, may influence free flow traffic speed. Traffic speeds for 690 vehicles traversing street blocks paved with asphalt, granite block, and brick materials were measured using a radar gun on a sample of 18 matched pair (asphalt and historic paving material) street blocks in Philadelphia, Pennsylvania. Fixed effects linear regression was used to estimate the effect of paving material on vehicle speed after controlling for the street class (e.g., arterial versus local road) and the matched pair. Results indicate that brick reduced speeds by approximately 3 mph (~5 km/h) and granite block reduced speeds by approximately 7 mph (~11 km/h), as compared to asphalt paved city streets, which we attribute to drivers intentionally slowing due to road roughness. This research suggests that brick and granite block paving materials may be an effective traffic calming strategy, having implications for reducing negative health outcomes associated with pedestrian–automobile collisions.
Dissipation of stop-and-go waves via control of autonomous vehicles: Field experimentsStern, RE; Cui, S; Delle Monache, ML; Bhadani, R; Bunting, M; Churchill, M; Hamilton, N; Haulcy, R; Pohlmann, H; Wu, F; Piccoli, B; Seibold, B; Sprinkle, J; Work, DB (2018-04-01)© 2018 Elsevier Ltd Traffic waves are phenomena that emerge when the vehicular density exceeds a critical threshold. Considering the presence of increasingly automated vehicles in the traffic stream, a number of research activities have focused on the influence of automated vehicles on the bulk traffic flow. In the present article, we demonstrate experimentally that intelligent control of an autonomous vehicle is able to dampen stop-and-go waves that can arise even in the absence of geometric or lane changing triggers. Precisely, our experiments on a circular track with more than 20 vehicles show that traffic waves emerge consistently, and that they can be dampened by controlling the velocity of a single vehicle in the flow. We compare metrics for velocity, braking events, and fuel economy across experiments. These experimental findings suggest a paradigm shift in traffic management: flow control will be possible via a few mobile actuators (less than 5%) long before a majority of vehicles have autonomous capabilities.