Attack Tolerance of RSS-based Fingerprinting

Abstract

This paper studies the performance of Received Signal Strength (RSS)-based<br/>fingerprinting positioning methods under different attack scenarios. We discuss<br/>different attack models and we compare the accuracy of a commonly used RSS<br/>fingerprinting algorithm with a robust version relying on access node<br/>visibility, with respect to those attacks. The results show that the robust<br/>fingerprinting method does not outperform the traditional fingerprinting method<br/>for most attacks, nevertheless, for a few attack types the accuracy<br/>improvement, in terms of Root Mean Square Error (RMSE), can yield factor two.<br/>RSS-based fingerprinting methods are most vulnerable to jamming, spoofing and<br/>interchanging of access nodes and least vulnerable to random removal of access<br/>nodes.<br/>