Experiencing and handling the diversity in data density and environmental locality in an indoor positioning service

Abstract

Diversity in training data density and environment locality is intrinsic in the real-world deployment of indoor localization systems and has a major impact on the performance of existing localization approaches. In this paper, through micro-benchmarks, we find that fingerprint-based approaches are preferable in scenarios where a dense database is available; while model-based approaches are the method of choice in the case of sparse data. It should be noted, however, that practical situations are complex. A single deployment often features both sparse and dense sampled areas. Furthermore, the internal layout affects the propagation of radio signals and exhibits environmental impacts. A certain number of measurement samples may be sufficient for one part of the building, but entirely insufficient for another. Thus, finding the right indoor localization algorithm for a given large-scale deployment is challenging, if not impossible; there is no one-size-fits-all indoor localization approach. Realizing the fundamental fact that the quality of the location database capturing the actual radio map dictates localization accuracy, in this paper, we propose Modellet, an algorithmic approach that optimally approximates the actual radio map by unifying model-based and fingerprint-based approaches. Modellet represents the radio map using a fingerprint-cloud that incorporates both measured real fingerprints and virtual fingerprints, which are computed from models with a local support, based on the key concept of the supporting set. We evaluate Modellet with data collected from an office building as well as 13 large-scale deployment venues (shopping malls and airports), located across China, U.S., and Germany. Comparing Modellet with two representative baseline approaches, RADAR and EZPerfect, demonstrates that Modellet effectively adapts to different data densities and environmental conditions, substantially outperforming existing approaches.