University of Twente Student Theses


Simulating Federated Learning for Smartphone based Indoor Localisation

Li, Litian (2021) Simulating Federated Learning for Smartphone based Indoor Localisation.

[img] PDF
Abstract:Satellite navigation such as Global Positioning System (GPS) cannot accurately and quickly locate indoors due to signal congestion and path complexity caused by the building structure. In indoor positioning technology based on wifi fingerprint is a general solution. As the demand for indoor positioning increases and people’s awareness of privacy protection increases. It is essential to protect privacy in the crowdsourced method of collecting user location information. Compared to traditional centralized machine learning and distributed machine learning. In federated learning, user data is only trained locally without leaving the local device. Only model parameters and gradients are transmitting between the service and the client. Thus federated learning has become a solution as a machine learning method to protect user privacy. The author will select a suitable open-source indoor positioning data set based on wifi fingerprints, and choose a suitable framework by evaluating the existing mainstream federated learning frameworks. Conducting federated learning and non-federal learning based on the selected data and framework. Observing whether it can have good training results under the premise of protecting user privacy characteristics of federated learning and compare it with the performance under traditional machine learning.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Computer Science BSc (56964)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page