Differential d-Vectors for RSS based Localization in Dynamic IoT Networks

Wireless Fidelity (Wi-Fi) Received Signal Strength (RSS) fingerprints are frequently used for localization in Internet of Things (IoT) environment. However, RSS fingerprinting based localization methods' accuracy degrades due to device heterogeneity and temporal variations (dynamic variations) in the target environment. In this paper, we propose a smart device localization method that addresses the dynamic RSS variations using differential d-vectors. The d-vector is the signature of a particular location and is persistent even with dynamic RSS variations. We compute the performance of the proposed d-vectors on two popular real-world datasets, and the obtained results outperform state-of-the-art fingerprinting methods that address the heterogeneity and temporal variations of the RSS values.