Graduate Symposium. Group C

Transcription

1 Graduate Symposium Group C BigData: Data Fusion-Big Data Coach: RINZIVILLO Salvatore Room number: C109 Day Session Time slot Presentation Mon 14/07 GS.1D 15:30 17:00 DUNNY Shane Tue 15/07 GS.2C 13:30 15:00 KAMGA Camille GS.2D 15:30 17:00 MADHAWA Kaushalya & MALDENIYA Danaja Thu 17/07 GS.4C a 13:30 14:15 SALANOVA GRAU Josep Maria b 14:15 15:00 LIU Feng GS.4D 15:30 17:00 VALERIO Danilo

2 DUNNY Shane AECOM, Ireland Interesting in the use of big data sources in the development of transportation and travel planning applications. Currently working on a fused data medium forecast model for use in an advance travel planning tool. The aim of the mode is to provide accurate forecasts of likely traffic scenarios ~12 hours in advance, which serve to inform road users who may leverage such information to make alternative choices. For operators, the proposed solution provides a passive travel demand management tool that can lead to better peak period spread should a fraction of users be influenced to change travel characteristics. This can significantly reduce emissions on roads, improve the reliability of the road network, reduce incidents and improve the overall efficiency of the road network. GS.1D Group C (BigData: Data Fusion Big Data)

3 KAMGA Camille University Transportation Research Center, USA I have interest on analyzing "big data" for transportation. Currently, my research focus on analyzing GPS data from vehicle probes and developing travel demand and prediction models. I am interested to use big data such as cell phone data for travel behavior analysis as well such as estimating O-D matrix and route choice models. GS.2C Group C (BigData: Data Fusion Big Data)

4 MADHAWA Kaushalya & MALDENIYA Danaja LIRNE Asia, Sri Lanka The primary research focus is to leverage Mobile Network Big Data or more accurately Transaction Generated Data (TGD) to gain insights on a large scale-mobility mobility, that can inform transportation planning in developing economies. We have obtained continuous access to historical and anonymized data from multiple mobile network operators in Sri Lanka covering more than 50 percent of the population (more than 10 million people). Our work involves the following: Identifying trip generation and attraction Development of Origin - Destination (OD) Matrices Predicting transportation mode and speed Identifying the daily mobility motif distribution in specific urban areas Predicting human activities conditionally based on public land use data supplemented with application APIs such a FourSquare, WikiMapia etc. Modelling Traffic flows in urban areas utilizing the out the outcomes of previous stages and road network data Modelling socio-economic characteristics of populations based mobility characteristics as well as Re-fill datasets Identifying the levels and nature of social interactions The challenges include unifying data from multiple operators accurately in the presence of anonymization, mapping cell tower based location data to fine grained administrative regions, compensating for sparse data and difficulties in validation due to scarcity of sources for getting ground truth values. At the present early stage of research preliminary work in identifying Home- Work hours, Work-Home locations at District Secretariat level, population density fluctuations during a day etc. has been done based on work by Sibren Isaacman et al.[1] Ongoing research includes trip generation, O-D matrices generation based on stay/stay region, virtual location concepts in work done by Gaston A. Fiore et al [2] as well transportation mode prediction. The research on estimating mode of transport is based on temporal distribution trips as introduced by H.Wang et al.[3] and Virtual Cell Path approach from J.Doyle et al s[4] work. [1] S. Isaacman, R. Becker, and R. Cáceres, Identifying important places in people s lives from cellular network data, Pervasive, pp. 1 18, [2] G. A. Fiore, Y. Yang, J. Ferreira, E. Frazzoli, and M. C. González, A Review of Urban Computing for Mobile Phone Traces : Current Methods, Challenges and Opportunities, [3] H. Wang, F. Calabrese, G. Di Lorenzo, and C. Ratti, Transportation mode inference from anonymized and aggregated mobile phone call detail records, in 13th International IEEE Conference on Intelligent Transportation Systems, 2010, pp [4] J. Doyle, P. Hung, D. Kelly, S. McLoone, and R. Farrell, Utilising mobile phone billing records for travel mode discovery, GS.2D Group C (BigData: Data Fusion Big Data)

5 SALANOVA GRAU Josep Maria Certh/Hit, Greece New technologies and social behavior have increased significantly the quantity and quality of mobility-related data available. The current challenge is twofold, on the one side there is a need for developing algorithms able to filter, validate and process vast amounts of data almost in real time, while on the other hand there is a need for developing new applications and services for providing innovative and advanced traveler information services based on theses new data and processing capabilities. I'm working during the last two years in two kinds of probe data sets: one collected by a network of more than 40 static sensors detecting Bluetooth MAC ids and a second one collected by a network of moving sensors (Floating Car Data) from a fleet of more than 1000 taxis. Both networks are installed and running in the city of Thessaloniki and are used for providing both information and routing services to drivers. Additionally, a set of cooperative services for both freight and passengers will be developed during the following 2-3 years, enriching the data sources in terms of quantity and quality. Various algorithms for data filtering, fusion and matching have been developed / are under development for processing, validating and using the collected data, which accounts for millions of daily records. GS.4C (1) Group C (BigData: Data Fusion Big Data)

6 LIU Feng UHasselt-IMOB, Belgium The world s urban population growth and economic development have led to the reshaping of metropolitan space layout among residential, employment and shopping locations, generating growing mismatch between travel demand and transport services. Although a variety of public policies have been introduced to ease the situation of the transport network, the measures are still lagging behind the pace of urban growth. A reliable method to accurately analyze the current mobility demand and underlying transport network systems as well as to identify the areas with serious mismatch problems, is thus important in the assistance of designing more effective measures. With the wide deployment of GPS devices in vehicles in many cities today, we explore the possibility of using GPS data to develop such an approach. Our exploration is composed of four major steps. First, city-wide mobility patterns are modeled based on GPS trajectories generated by vehicles. This model captures a set of key traffic characteristics between each pair of regions of the entire city network, including travel demand, travel speed, and route directness of travel paths. Upon this model, a set of indicators is then built to measure the road transport performance between the regions, and the areas with serious mismatch problems are subsequently pinpointed. Finally, the identified problematic regions are further examined and specific transport problems are analyzed. By applying the proposed method to the Chinese city of Harbin using GPS data collected from all taxis operating in the city between July and September in 2013, the potentials and effectiveness of this technique are demonstrated. With more and more urban vehicles being installed with GPS devices, the designed method can be easily transferrable to the cities, thus paving a way for the development of a new, up-to-date and spatial-temporal sensitive road network analysis approach that supports the urban growth and transport system development into a sustainable future. GS.4C (2) Group C (BigData: Data Fusion Big Data)

7 VALERIO Danilo FTW Telecommunications research Centre, Austria Collecting extensive information about human and vehicular mobility is a fundamental prerequisite for intelligent transportation system (ITS) applications. Traditional approaches to collect mobility data and traffic status information are prone to several technical and economical limitations. Systems based on road traffic detectors and cameras suffer from high installation costs, which pose an obstacle to the extensive coverage of a road network. Systems based on floating car data (FCD) are limited by the size and representativeness of specific sub-sets of probes. Finally, census and questionnaires (which are still the main source of large-scale mobility information and commuting statistics) are costly, time consuming, and inaccurate. The recent spread of wireless technologies provides unprecedented possibilities for the collection of mobility data. My PhD focuses on the extraction and analysis of anonymous mobility traces from a large-scale operational cellular network. The goal is to design a system that is able to process cellular signalling data of a mobile operator in order to (a) extract mobility trajectories, (b) compute real-time measurements, and (c) extract mobility patterns and statistics from historical data. Real-time applications include, e.g., inference of current traffic intensity, detection of traffic congestions, and estimation of expected travel times. Historical data analysis instead aims at studying human mobility patterns in order to provide support for tasks like road infrastructure planning, urban planning, optimization of public transport, etc. At the time of writing, I tackled most of the aspects related to real-time traffic monitoring. I developed a system that, based on real-world cellular network data from a mobile operator, accurately estimates travel times in highways and, in turn, generate road congestion alarms based on fluctuations of the estimated travel times. In addition, I developed a module that triggers alarm whenever an anomaly in the cellular network signalling is likely to be caused by an anomaly in vehicular traffic. Recently, I started working on the analysis of cellular data for the extraction of mobility patterns. The goal here is to extract O/D matrices and monitor multi-modal trips. Given my telecommunications background, I decided to pay a visit to a PhD school from the Transportation research community. I hope to grasp as much as possible about new trends in mobility patterns, mobility modelling, multi-modal trips, and big data for mobility analysis. GS.4D Group C (BigData: Data Fusion Big Data)