Toward secure and usable indoor positioning
Date
2022
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
University of Delaware
Abstract
Recent years have witnessed growing interests from both academia and industry in developing effective Indoor Positioning Systems (IPSes). An IPS allows users to learn their locations and navigate in large unfamiliar indoor venues such as shopping malls, hospitals, and airports, where GPS signals are absent or unreliable. Despite significant efforts made in the past, existing IPSes suffer from two major limitations. First, most existing IPSes are designed for improving positioning accuracy without considering potential malicious attacks. Second, many IPSes require costly infrastructure upgrades, which may hinder their wide adoption and deployment. This dissertation tackles three challenges in developing secure and usable IPSes to pave the way for their wide adoption. ☐ First, we introduce MV-IPS, a novel WiFi-based IPS that can achieve high positioning accuracy regardless of the presence of malicious attacks. It is well known that a malicious attacker can severely degrade the positioning accuracy of WiFi-based IPSes by deploying fake WiFi Access Points (APs) to impersonate legitimate ones. We observe that existing WiFi-based IPSes suffers from an inherent trade-off between positioning accuracy and attack resilience, i.e., the strong attack resilience comes at the cost of reduced positioning accuracy in the absence of attacks. To tackle this challenge, we develop a novel multi-vote WiFi-based IPS without suffering from such trade-off. In the proposed IPS, we view APs as voters and determine a user's location by jointly considering APs' weighted votes. Doing so can effectively limit the impact of potential malicious APs without sacrificing the positioning accuracy in the absence of attacks. ☐ Second, we introduce Impos, a novel image-based IPS for indoor venues where WiFi infrastructure is unavailable, which allows a user to learn her location by taking a small number of photos of the surrounding using smartphones. While several image-based IPSes have been proposed in the literature, they all suffer from low positioning accuracy due to inaccurate angle measurements from smartphone sensor readings. To address their limitation, our system greatly improves the angle measurement through image analysis and effectively copes with varying number of landmarks. Detailed experimental studies confirm the significant advantages of our IPS over prior solutions. ☐ Last but not the least, we develop CITS, a novel RSS-based continuous indoor tracking system. Existing RSS-based IPSes suffer from the ambiguity of RSS fingerprints and device homogeneity which could greatly limit its positioning accuracy in practice. To tackle these challenges, CITS estimates a user's location via differential RSS fingerprint matching and path tracking. Our experiment results confirm that CITS can achieve high positioning accuracy in the presence of fingerprint ambiguity and device diversity and significantly outperform prior solutions.
Description
Keywords
Indoor positioning system, WiFi, Positioning accuracy, Attack resilience