Localizing Heterogeneous Access Points using Similarity-based Sequence

Abstract

Positioning of Wifi access points (APs) is important to understand the nature of the deployed IEEE 802.11 network, for example, coverage, connectivity, and density. More importantly, using Wifi to perform indoor localization is a promising solution, and understanding the Wifi AP location could potentially improve the indoor localization accuracy. A common approach is to use an explicit model, which describes the signal propagation over a distance. But predicting the signal propagation is challenging due to the AP diversity, different antenna gains, occlusion and multi-path issue in indoor environments. Therefore, we propose to use sequence-based approach for the localization of heterogeneous APs. In particular, we represent a position with location sequence and the measurement with RSS (received signal strength) sequence. The localization of an AP is achieved by measuring the similarity between the location and RSS sequence. Moreover, we design a new similarity measure that considers the quality of a match to improve the localization accuracy. The proposed approach can be used to localize diverse APs under different antenna propagation characteristics without modeling the propagation of the radio signal. Experiments were carried in an indoor environment with 8 APs and our results show that our approach outperforms the propagation model-based approach and the sequence-based approach by 32.2% and 19.5%, respectively.