In indoor localization our task is to use wireless signals impinging on a multi-antenna array to determine the location of a mobile device. Our approach is fingerprinting, that is - we obtain a dataset of signals sent from various locations in our desired area. From these signals, we extract relevant features such as direction and time of arrival. The location of an incoming signal is determined by comparing its features to the dataset.

​

Our first paper describes the basic method for determining location:
Paper: Single-Site Localization via Maximum Discrimination Multipath Fingerprinting

​

Our second paper's main topic is a theoretical analysis for a semi-supervised algorithm: Geodesic nearest neighbours. Our main application for the algorithm is indoor localization. 

Paper: Minimax-optimal semi-supervised regression on unknown manifolds

​

Indoor localization