Efficient Indoor Localization with Multiple Consecutive Geomagnetic Sequences

Abstract

Geomagnetism-based indoor localization has great social and commercial value due to its pervasiveness and indepen-dence from extra infrastructure. To improve the distinguishability of geomagnetic signals as location clues, geomagnetic sequences are usually taken as input. Although longer input sequence can provide higher localization accuracy, it suffers from high response time in practice. To address the above, we first utilize short geomagnetic sequences as input, alleviating high response time, and propose an efficient single position estimation model, taking advantage of modified transformer to estimate position for each independent short sequence. Noticing the temporal dependency and the spatial consistency constraint during continuous positioning, we further propose a joint position estimation model to capture the correlations among consecutive short sequences, achieving higher accuracy with multiple short sequences. We have conducted extensive experiments in two typical trial sites, a narrow office area and a spacious parking lot. Experimental results show that the proposed approach outperforms state-of-the-art competing schemes, and the localization error is reduced by more than 32% with shorter geomagnetic sequences.