TOP
Search the Dagstuhl Website
Looking for information on the websites of the individual seminars? - Then please:
Not found what you are looking for? - Some of our services have separate websites, each with its own search option. Please check the following list:
Schloss Dagstuhl - LZI - Logo
Schloss Dagstuhl Services
Seminars
Within this website:
External resources:
  • DOOR (for registering your stay at Dagstuhl)
  • DOSA (for proposing future Dagstuhl Seminars or Dagstuhl Perspectives Workshops)
Publishing
Within this website:
External resources:
dblp
Within this website:
External resources:
  • the dblp Computer Science Bibliography


Dagstuhl Seminar 22162

Urban Mobility Analytics

( Apr 18 – Apr 22, 2022 )

(Click in the middle of the image to enlarge)

Permalink
Please use the following short url to reference this page: https://www.dagstuhl.de/22162

Organizers

Coordinator

Contact



Schedule

Summry

Seminar 22162 addressed recent trends in urban mobility analytics that are shaping the information available to transport planners, operators, and travellers. Seminar participants were particularly discussing how information can be provided that supports the critical transformation of urban mobility towards climate neutrality and other sustainability goals, i.e, that supports to change mobility behaviour.

The trends identified for this seminar were, on one hand, the rise of deep learning methods for massive data analytics, and on the other hand the emerging digital divide between those having massive data and those who haven't, which, in short, forms the challenges of academia for reproducible research. Massive data on urban mobility is collected by industry and transport authorities, with limited access outside, for various reasons. Also, the research and development capacity behind the closed doors of large transnational companies -- especially in the platform economy -- is arguably faster than the typical PhD process.

These challenges and opportunities were shaping the discussions where participants split into working groups on (a) ethics and the social good -- how can information trigger change in mobility behaviour; (b) methods and explainability; (c) benchmarking and datasets; and (d) applications.

The seminar had quite a diversity of participants, which was inspiring in all the discussions. Participants from industry gave talks about what happened behind their "closed doors', and further tutorials were introducing datasets, the principle of reproducible research, and European funding opportunities.

The industry partners showed great interest in collaboration with academia, however, the problem of data sharing was still considered as paramount. There are trends to open certain kinds of data, e.g. in aggregated form, or simulated data, or only based on contracts with certain institutions. Still, open data sharing remains to be a challenge.


Motivation

The Dagstuhl Seminar aims to bring together researchers from academia and industry who work in complementary ways on urban mobility analytics such that they do not necessarily meet at the same conferences. Especially we aim to collide ideas and approaches from deep learning research – requiring large datasets – and reproducible research – requiring access to data.

Transportation in cities is undergoing unprecedented change – such as by vehicle technology towards autonomous driving (‘disruptive mobility’); massive real-time data and data analytics (smart cities, sensing cities, dashboards); sharing platforms and integration (mobility-as-a-service) and urban logistics (changing shopping patterns). All this happens in parallel with an increasing willingness to share mobility resources or change mobility behaviour. Critical to the success of transforming urban mobility is therefore information provided to planners, operators, and travellers.

The seminar will address recent trends that are shaping the information derived from urban mobility analytics.

  • A prominent one, not only in transportation research, is the rise of deep learning methods for massive data analytics. In the domain of urban mobility this massive data emerges from a range of sensor platforms, from infrastructure (CCTV, induction loops, people counters, WiFi, smart cards, air quality) to vehicles (GPS, vision, LiDAR, radar) and smartphones (GPS, location-based apps, accelerometer, gyroscope, magnetometer), in volume, heterogeneity, velocity and veracity a prime application domain for deep learning.
  • A second trend is the emerging digital divide between academia and industry and its challenges for reproducible research, a trend that has been compared to digital feudalism. While massive data on urban mobility is collected by industry and transport authorities, their access for academic research is limited by privacy concerns and also by commercial sensitivities. While reproducible research hinges on access to data, much of the urban mobility research and development is now done behind the closed doors of large transnational companies.

Related to both trends above is the buzzword of Digital Twins. Since both data and data analytics becomes more often available sufficiently close to real-time, the information derived is less and less consumed in human decision making but in the self-regulation of cyber-physical-social systems. These systems will propel future urban mobility by autonomously driving vehicles and mobility-as-a-service, however, their development and use involves many still-open research questions, such as reliability, trust and HCI.

In this context the seminar makes a deliberate effort to invite people from both sides of the digital divide (i.e., from academia and industry) to share their experiences, their approaches, and their challenges, and to explore more future collaboration.

We aim for a joint vision paper of the participants as much as new international collaborative research initiatives. The dynamics at the seminar will be fostered by a large traffic dataset sponsored by IARAI (derived from https://www.iarai.ac.at/traffic4cast/). This dataset will be available for the working groups to explore and test their ideas.

Copyright David Jonietz, Monika Sester, Kathleen Stewart, and Stephan Winter

Participants
On-site
  • Vanessa Brum-Bastos (Wroclaw University of Environmental and Life Sci., PL)
  • Hao Cheng (Leibniz Universität Hannover, DE)
  • Tao Cheng (University College London, GB)
  • David Doerr (TÜV Rheinland - Köln, DE)
  • Christian Eichenberger (IARAI - Zürich, CH)
  • Cheng Fu (Universität Zürich, CH)
  • Martin Lauer (KIT - Karlsruher Institut für Technologie, DE)
  • Dirk Christian Mattfeld (TU Braunschweig, DE) [dblp]
  • Alexandra Millonig (AIT - Austrian Institute of Technology - Wien, AT)
  • Edoardo Neerhut (Meta - Burlingame, US)
  • Moritz Neun (IARAI - Zürich, CH)
  • Daniel Nüst (Universität Münster, DE)
  • Erik Nygren (Schweizerische Bundesbahnen - Bern, CH)
  • Maya Sekeran (TU München, DE)
  • Monika Sester (Leibniz Universität Hannover, DE) [dblp]
  • Martin Tomko (University of Melbourne - Carlton, AU)
  • Stephan Winter (The University of Melbourne, AU) [dblp]
  • Yanan Xin (ETH Zürich, CH)
Remote:
  • Geoff Boeing (USC - Los Angeles , US)
  • Andris Clio (Georgia Institute of Technology - Atlanta, US)
  • Ioannis Giannopoulos (TU Wien, AT)
  • Anita Graser (AIT - Austrian Institute of Technology - Wien, AT)
  • David Jonietz (HERE - Zürich, CH) [dblp]
  • Ivan Majic (TU Graz, AT)
  • Harvey J. Miller (Ohio State University, US) [dblp]
  • Michael Nolting (Volkswagen Nutzfahrzeuge - Hannover, DE)
  • Luca Pappalardo (CNR - Pisa, IT)
  • Chiara Renso (ISTI-CNR - Pisa, IT) [dblp]
  • Kathleen Stewart (University of Maryland - College Park, US) [dblp]
  • Piyushimita Vonu Thakuriah (Rutgers University - New Brunswick, US) [dblp]

Related Seminars
  • Dagstuhl Seminar 10121: Computational Transportation Science (2010-03-21 - 2010-03-26) (Details)
  • Dagstuhl Seminar 13512: Social Issues in Computational Transportation Science (2013-12-15 - 2013-12-19) (Details)
  • Dagstuhl Seminar 16091: Computational Challenges in Cooperative Intelligent Urban Transport (2016-02-28 - 2016-03-04) (Details)

Classification
  • Artificial Intelligence
  • Machine Learning
  • Multiagent Systems

Keywords
  • Machine learning
  • deep learning
  • massive data analytics
  • travel prediction