Doctoral School : Sciences et Technologies de l'Information et de la Communication and the Research Unit SAMOVAR (UMR 5157) - Services répartis, Architectures, Modélisation, Validation, Administration des Réseaux are presenting the "examination of a thesis" by Mr Yasir Saleem SHAIKH who is expected to defend his research to obtain his PhD at l'Institut Polytechnique de Paris, prepared at Télécom SudParis in : Computer science
« Protection de la Confidentialité des Services de Recommandation pour les Villes Intelligentes »
On Tuesday , January 21, 2020 à 9h30 - at Télécom SudParis - 9 rue Charles Fourier - 91000 Évry
Jury membres :
- M. Noel CRESPI, Professor, Télécom SudParis, FRANCE - Supervisor
- Mme Maria POTOP-BUTUCARU, Professor, Sorbonne Université, FRANCE - Examiner
- M. Luigi ATZORI, Professor, University of Cagliari, ITALY - Examiner
- M. Martin BAUER, Researcher, NEC Europe Ltd, GERMANY - Examiner
- M. Roberto MINERVA, Assistant professor, Télécom SudParis, FRANCE - Supervisor
- M. Payam BARNAGHI, Professor, University of Surrey, UK - Reviewer
- M. JaeSeung SONG, Associate professor, Sejong University, KOREA - Reviewer
During the past decade, the Internet of Things (IoT) technology has revolutionized almost all the fields of daily life and has boosted smart cities. Smart cities use IoT technology to collect various types of sensors’ data and then use such data to offer a variety of applications. Since the smart cities’ applications are used by the citizens, therefore providing the customized recommendation services to the citizens based on their preferences, locations and profiles, as well as by exploiting the IoT data (e.g., traffic congestion and parking occupancy) is of great importance which could be provided by an IoT recommender. However, since the IoT recommender utilizes the private data of citizens (e.g., profiles, preferences and habits), it breaches the privacy of the users because the IoT recommender could track the routines and habits of the users by analyzing the historical database or by analyzing the regular recommendation services it offers.
Therefore, it is important to preserve the privacy of the users from the IoT recommender. In this thesis, we propose a novel privacy preserving IoT recommender system for smart cities that provides recommendations by exploiting the IoT data of sensors and by considering various metrics. Our approach is organized in three parts. Firstly, we develop an EU General Data Protection Regulation (GDPR)-compliant IoT recommender system for smart parking system that provides recommendations of parking spots and routes by exploiting the data of parking and traffic sensors. For this, we first propose an approach for the mapping of traffic sensors with route coordinates in order to analyze the traffic conditions (e.g., the level of congestion) on the roadways and then developed an IoT recommender. The IoT recommender has been integrated into the smart parking use case of an H2020 EU-KR WISE-IoT project and has been evaluated by the citizens of Santander, Spain through a prototype.
Additionally, we develop an IoT recommender for smart skiing that provides skiing routes comprised of specific types of slopes, as well as the nearest slope. For skiing routes, there does not exist any stable routing engine. Therefore, a novel routing engine for skiing routes was developed. This work has also been integrated into the smart skiing use case of WISE-IoT project. Secondly, although the developed IoT recommender for smart parking is GDPR-compliant, however, it does not fully protect the privacy of users. Because, an indiscriminately sharing of users’ data with untrusted third-party IoT parking recommender system causes a breach of privacy, as user’s behavior and mobility patterns can be inferred by analyzing the past travelling history of users. Therefore, we preserve privacy of users against parking recommender system while analyzing their past parking history using k-anonymity and differential privacy techniques.
Lastly, since the smart cities applications are developed in a vertical manner and do not talk/communicate with each other, i.e., each application is developed for a certain scenario which generally does not share data with other smart cities applications. Therefore, we proposed two frameworks for the recommendation services across smart cities applications using social IoT. Firstly, on how social IoT can be used for the recommendation services across smart cities applications, and secondly, we propose another type of communication of social IoT at a global level, i.e., social cross-domain application-to-application communications, that enables smart cities applications to communicate with each other and establish social relationships between them.