Back to overview

MOBIDICT - Privacy-Aware Predictive Platform for Green Mobility

English title MOBIDICT - Privacy-Aware Predictive Platform for Green Mobility
Applicant Garbinato Benoit
Number 157160
Funding scheme Project funding (Div. I-III)
Research institution Département des systèmes d'information Faculté des Hautes Etudes Commerciales Université de Lausanne
Institution of higher education University of Lausanne - LA
Main discipline Information Technology
Start/End 01.11.2015 - 31.10.2019
Approved amount 238'966.00
Show all

Keywords (5)

information dissemination and aggregation; distributes algorithms; sofware engineering; green transportation services; mobile ad hoc networks

Lay Summary (French)

La mobilité individuelle a considérablement augmenté au cours des dernières décennies grâce à l'omniprésence des véhicules privées, ce qui a permis à une partie croissante de la population de déménager à la campagne, dans l'espoir de vivre plus sainement. Paradoxalement, cette augmentation de la mobilité a également créé de graves problèmes environnementaux. Pour y faire face, la tendance actuelle consiste à pousser les gens vers des alternatives plus durables, comme les transports publics ou le covoiturage. Malheureusement, ces alternatives ont aussi leurs limites, car elles imposent de fortes contraintes de planification. De plus, les transports publics ne sont pas toujours économiquement viables. Une approche prometteuse ici consiste à combiner transports publics et transports privés, en s'appuyant les technologies mobiles. Un défi clé ici réside dans la nécessité de prédire la mobilité des utilisateurs, tout en préservant leur vie privée.
Lay summary

Contenu et objectifs du travail de recherche

Ce projet vise à explorer l'antagonisme naturel existant entre prédiction de la mobilité des utilisateurs d'une part, et respect de leur vie privée d'autre part, dans le cadre d'une plate-forme encourageant la mobilité verte et accessible via les technologies mobiles. Cette plate-forme se appuiera sur trois services: un moteur de prédiction de la mobilité (mobility prédiction), un service de calcul réparti sécurisé (secure multiparty computation) et un service de publication/souscription basé sur la localisation (location-based publish/subscribe). Le moteur de prédiction sera chargé de prévoir les futures localisations des utilisateurs, en fonction de leurs habitudes de mobilité. Le moteur sera basé sur deux composants : un moteur local intégré dans chaque appareil mobile et un moteur global disponible à tous les appareils. Le compromis entre prédiction et vie privée dépendra des rôles respectifs joués par le moteur global et les moteurs locaux.  Le service de calcul réparti coordonnera les interactions entre appareils mobiles, en assurant leur confidentialité, et le service de publication/souscription assura l'acheminement asynchrone des messages en se basant sur la localisation des appareils.

Contexte scientifique et social du projet de recherche

Ce projet permettra de mieux comprendre l'antagonisme existant entre prédiction de la mobilité et respect de la vie privée, en proposant une mesure précise de cet antagonisme. Cette mesure permettra de comparer quantitativement les divers compromis possibles entre prédiction de la mobilité et respect de la vie privée, notamment lors de l'élaboration de nouveaux algorithmes de prédiction dans le cadre de notre plate-forme encourageant la mobilité verte. Afin de valider notre approche, nous prévoyons de tester cette plate-forme sur une base d'utilisateurs fournie par la Section Vaud de l'Association Transports et Environnement (ATE).

Direct link to Lay Summary Last update: 09.12.2014

Responsible applicant and co-applicants


Name Institute

Associated projects

Number Title Start Funding scheme
138092 MAHGA-2 : Mobility Models and Communication Patterns of Mobile Ad Hoc Gaggles in Wireless Mesh Networks 01.10.2011 Project funding (Div. I-III)
140762 SOFTEGE: Adaptive Software Platform for Context-aware Green Transportation Services 01.10.2012 Project funding (Div. I-III)


Context. Individual mobility has dramatically increased over the past decades, in particular thanks to the growing ubiquity of private motor vehicles as flexible transport means. As a consequence, a significant portion of the population relocated to peri-urban areas, if not to the countryside, in the hope to live in a healthier environment. Paradoxically, this increase in fuel-based mobility has created serious environmental issues. To address them, the current trend consists in pushing people to shift from individual transport to more sustainable alternatives. Public transport and car sharing are two such alternatives, but their scope tends to be limited, partly because they impose strong planning constraints on both individuals and public authorities, and partly because public transport are simply not economically viable in certain remote locations.Problem. For journeys no viable by public transport only, a promising approach consists in trying to dynamically combine public and private transport. The goal is to reach the same flexibility offered by private motor vehicles. This approach is part of a new research stream known as Green Information Systems (Green IS), which aims at devising information systems that foster sustainable behaviors. While ambitious, the idea of a Green IS dynamically combining public and private transport is gradually becoming realistic, thanks to the rapid development of mobile technologies. Yet key challenges need to be addressed before such information systems can be implemented. A first challenge lies in the heterogeneity of its components: various mobile devices (smartphones, tablets, car navigation systems, etc.), various communication protocols (LTE, WiFi, Bluetooth, ad hoc networks, etc.), various information sources (public transport timetables, car sharing availability, parking lot booking, etc.). This heterogeneity requires the system to be extensible to continuously integrate new information sources, new protocols, new devices, etc. A second challenge lies in the need to predict the state of those components in real time, in particular the mobility of users. Interestingly, besides transport-oriented resources, which are easy to monitor and to predict, human mobility also seems to be predictable in most cases. However, predicting the mobility of people comes with an inherent tradeoff between accuracy and privacy.Approach. This project will explore the tradeoff between accuracy and privacy, by building experimental predictive platform that encourages green mobility while preserving the privacy of its users. At its core, the platform will rely on three basic services: a mobility prediction engine, a privacy-aware multiparty computation service, and a location-based publish/subscribe service. The mobility prediction engine will be responsible for predicting the future locations of mobile users, based on their mobility patterns. The engine will actually be composed of two parts: a local engine embedded in each mobile device and a global engine available to all devices. The tradeoff between accuracy and privacy will be reflected in the role played by the global engine relative to the local engines. The privacy-aware multiparty computation service will then support the coordination between mobile devices, while ensuring fairness and privacy, whereas the location-based publish/subscribe service will be used for anonymous and asynchronous communications. It is important to note that these three core services should be able to operate not only in traditional wireless networks but also in mobile ad hoc networks (MANETs). Finally, to address the problem of heterogeneous information sources, the platform will rely on a plugin architecture, where each source is accessed via a plugin that conforms to a generic and open Service Provider Interface (SPI).Validation. To validate our approach, we will test the platform on a user base provided by the Vaud Chapter of the Association Transport et Environnement (ATE). The ATE had already agreed to participate in a similar experiment in the context of the Softege project (SNF 140762), but the experiment could not be carried out due to cuts in the requested resources. While the two projects share the same high-level goal (promoting green mobility), they differ in their approach and focus. As a reminder, the initial Softege project aimed at providing a software stack to developers building Green IS, and required financing for two PhD students and one PostDoc position. Since only one PhD student was funded, the scope of Softege was narrowed to the core layers of the stack, which are essentially communication-oriented. The good news is that this project will benefit from the agreement of the ATE to test the platform and from the results of the Softege project. Furthermore, we intend to promote to resulting platform as an open-source and extensible project, in order to disseminate the results of this research.