Broad Learning System for Indoor CSI Fingerprint Localization

With the development of the Internet of Things (IoT), the demand for location-based services in indoor environments proliferates. Channel State Information (CSI) based fingerprint localization has become a research hot spot. However, a compelling method to address this issue does not appear because of the hardship of processing CSI data. Meanwhile, many fingerprint localization algorithms have a time-consuming offline training phase. Therefore, we propose an indoor CSI fingerprint localization system based on the broad learning system (BLS). First, we filter the outliers, generate delegates of CSI by the data nugget algorithm, and use tensor decomposition to reconstruct the CSI delegates. Moreover, we utilize Isometric mapping (Isomap) to extract CSI features to reduce BLS's complexity. The experimental results show that our scheme outperforms several existing algorithms in two indoor environments.