An INDOOR LOCATION-BASED AUTHENTICATION SYSTEM (ILBAS): Indoor Location Detection and Authentication of Smartphones based on Received Wi-Fi Signal Strength

Abstract

The pervasiveness of location-based services (LBS) has increased the need for authentication systems that use location data as an authentication factor. Most of the existing systems are limited to outdoor locations, require special equipment, or do not have sufficiently high location detection accuracy. In this project we exploit the widespread use of the Wi-Fi access points (APs) and the ability of smartphones to receive wireless signals in detecting the smartphone location. We propose an indoor location-based authentication system (ILBAS) for smartphones based on received Wi-Fi signal strength using machine learning algorithms (MLAs). Three common algorithms are used to evaluate the effectiveness of Wi-Fi fingerprints: support vector machine (SVM), k-nearest neighbour (KNN), and deep learning. The ILBAS system includes filtration and pre-processing of the Wi-Fi fingerprint before the machine learning algorithm is applied. Different filtration mechanisms result in different fingerprints, which in turn show a different performance on each MLA. The SVM and KNN algorithms showed the best performance, and the fingerprints with the smallest number of AP exhibited the highest detection accuracies. The ILBAS system has been designed in such a way that it can be added to any existing authentication system. In this dissertation we design a two- factor authentication protocol. This protocol suggests using ILBAS along with another authentication factor, the claimant’s private key, to show its applicability and improve its security.