Ioulia Markou

Localization and Driving Behavior Classification with Smartphone Sensors in Direct Absence of Global Navigation Satellite Systems

Global navigation satellite systems have tremendous impact and potential in the development of intelligent transportation systems and mobility services and are expected to deliver significant benefits, including increased capacity, improved safety, and decreased pollution. However, there are situations in which there might not be direct location information about vehicles, for example, in tunnels and in indoor facilities such as parking garages and commercial vehicle depots. Various technologies can be used for vehicle localization in these cases, and other sensors that are currently available in most modern smartphones, such as accelerometers and gyroscopes, can be used to obtain information directly about the driving patterns of individual drivers. The objective of this research is to present a framework for vehicle localization and modeling of driving behavior in indoor facilities or, more generally, facilities in which global navigation satellite system information is not available. Localization technologies and needs are surveyed and the adopted methodology is described. The case studies, which use data from multiple types of sensors (including accelerometers and gyroscopes from two smartphone platforms as well as two reference platforms), provide evidence that the opportunistic smart-phone sensors can be useful in identifying obstacles (e.g., speed humps) and maneuvers (e.g., U-turns and sharp turns). These data, when cross-referenced with a digital map of the facility, can be useful in positioning the vehicles in indoor environments. At a more macroscopic level, a methodology is presented and applied to determine the optimal number of clusters for the drivers’ behavior with a mix of suitable indexes.

Towards distribution-based calibration for traffic simulation

Traffic simulation models have seen increasing use during the past decades. One of the biggest challenges related to their successful application is effective calibration and validation. Emerging data collection techniques provide richer data that can be used to improve this process. In this research, we explore the use of distributions of collected data (such as accelerations, using opportunistic sensors, such as smart-phone accelerometers) for calibration purposes. The performance of the considered ubiquitous sensors is benchmarked against reference equipment, to evaluate its accuracy under different conditions. A methodology is proposed for the integration of distributions of data in traffic simulation model calibration and validation.

A demonstration of distribution-based calibration

Calibration plays a fundamental role in successful applications of traffic simulation models and Intelligent Transportation Systems. In this research, the use of distributions in calibration process is motivated. The optimization of model parameters is fulfilled using the Simultaneous Perturbation Stochastic Approximation (SPSA) algorithm. The output of the optimization is a distribution of parameter values, capturing a wide range of various traffic conditions. As a proof of concept, a case study is also presented where the proposed framework is implemented for the distribution-based calibration of the car-following model used in the TransModeler microscopic traffic simulation model. The use of parameter distributions is preferred to using point parameter values, as it is more realistic, capturing the heterogeneity of driver behavior, and allows the simultaneous study of various driving behavior patterns. Flexibility is thus introduced into the calibration process and restrictions generated by conventional calibration methods are relaxed.

Combining time-series and textual data for taxi demand prediction in event areas: a deep learning approach.

Accurate time-series forecasting is vital for numerous areas of application such as transportation, energy, finance, economics, etc. However, while modern techniques are able to explore large sets of temporal data to build forecasting models, they typically neglect valuable information that is often available under the form of unstructured text. Although this data is in a radically different format, it often contains contextual explanations for many of the patterns that are observed in the temporal data. In this paper, we propose two deep learning architectures that leverage word embeddings, convolutional layers and attention mechanisms for combining text information with time-series data. We apply these approaches for the problem of taxi demand forecasting in event areas. Using publicly available taxi data from New York, we empirically show that by fusing these two complementary cross-modal sources of information, the proposed models are able to significantly reduce the error in the forecasts.

Use of Taxi-Trip Data in Analysis of Demand Patterns for Detection and Explanation of Anomalies

Because of environmental and economic stress, current strong investment in adaptive transport systems can efficiently use capacity, minimizing costs and environmental impacts. The common vision is of a system that dynamically changes itself (the supply) to anticipate the needs of travelers (the demand). In some occasions, unexpected and unwanted demand patterns are noticed in the traffic network; these patterns lead to system failures and cost implications. Significantly, low speeds or excessively low flows at an unforeseeable time are only some of the phenomena that are often noticed and need to be explained for a transport system to develop a better future response. The objective of this research was the formulation of a methodology that could identify anomalies on traffic networks and correlate them with special events by using Internet data. The main subject of interest in this study was the investigation of why traffic congestion was occurring as well as why demand fluctuated on days when there were no apparent reasons for such phenomena. The system was evaluated by using Google’s public data set for taxi trips in New York City. A “normality” baseline was defined at the outset and then used in the subsequent study of the demand patterns of individual days to detect outliers. With the use of this approach it was possible to detect fluctuations in demand and to analyze and correlate them with disruptive event scenarios such as extreme weather conditions, public holidays, religious festivities, and parades. Kernel density analysis was used so that the affected areas, as well as the significance of the observed differences compared with the average day, could be depicted.