SENTINEL: Securing Indoor Localization against Adversarial Attacks with Capsule Neural Networks

TL;DR

SENTINEL is a capsule neural network-based framework that significantly enhances indoor localization robustness against adversarial attacks, device heterogeneity, and RSS fluctuations, validated on a new rogue AP dataset.

Contribution

The paper introduces SENTINEL, a novel embedded ML framework using modified capsule neural networks to improve indoor localization security and accuracy against adversarial threats and environmental variability.

Findings

Up to 3.5x reduction in mean localization error.

Up to 3.4x reduction in worst-case error under simulated attacks.

Up to 2.8x improvement in mean error on real-world dataset.

Abstract

With the increasing demand for edge device powered location-based services in indoor environments, Wi-Fi received signal strength (RSS) fingerprinting has become popular, given the unavailability of GPS indoors. However, achieving robust and efficient indoor localization faces several challenges, due to RSS fluctuations from dynamic changes in indoor environments and heterogeneity of edge devices, leading to diminished localization accuracy. While advances in machine learning (ML) have shown promise in mitigating these phenomena, it remains an open problem. Additionally, emerging threats from adversarial attacks on ML-enhanced indoor localization systems, especially those introduced by malicious or rogue access points (APs), can deceive ML models to further increase localization errors. To address these challenges, we present SENTINEL, a novel embedded ML framework utilizing modified capsule neural networks to bolster the resilience of indoor localization solutions against adversarial attacks, device heterogeneity, and dynamic RSS fluctuations. We also introduce RSSRogueLoc, a novel dataset capturing the effects of rogue APs from several real-world indoor environments. Experimental evaluations demonstrate that SENTINEL achieves significant improvements, with up to 3.5x reduction in mean error and 3.4x reduction in worst-case error compared to state-of-the-art frameworks using simulated adversarial attacks. SENTINEL also achieves improvements of up to 2.8x in mean error and 2.7x in worst-case error compared to state-of-the-art frameworks when evaluated with the real-world RSSRogueLoc dataset.