Radio Map Crowdsourcing Update Method Using Sparse Representation and Low Rank Matrix Recovery for WLAN Indoor Positioning System
Yongliang Zhang, Lin Ma
Abstract
Updating radio map quickly and accurately is very challenging when the crowdsourcing radio fingerprints is provided by the unprofessional volunteers in WLAN indoor positioning system. To solve the problem, we propose a sparse representation and low rank matrix recovery based radio map update method. This method uses fingerprint correlation learned by sparse representation to complete the radio map consisting of fingerprint patches. Furthermore, the low-rank and sparse prior are combined skillfully in our proposed method to handle the fingerprint missing and sparse noise. Based on our analysis and experimental results, the proposed method significantly outperforms the state-of-the-art radio map update method even when the available crowdsourcing data accounts for a low degree of the entire radio map volume.