Large-Scale Location-Aware Services in Access: Hierarchical Building/Floor Classification and Location Estimation using Wi-Fi Fingerprinting Based on Deep Neural Networks

摘要

One of key technologies for future large-scale location-aware services inaccess is a scalable indoor localization technique. In this paper, we reportpreliminary results from our investigation on the use of deep neural networks(DNNs) for hierarchical building/floor classification and floor-level locationestimation based on Wi-Fi fingerprinting, which we carried out as part of afeasibility study project on Xi'an Jiaotong-Liverpool University (XJTLU) CampusInformation and Visitor Service System. To take into account the hierarchicalnature of the building/floor classification problem, we propose a new DNNarchitecture based on a stacked autoencoder for the reduction of feature spacedimension and a feed-forward classifier for multi-label classification withargmax functions to convert multi-label classification results into multi-classclassification ones. We also describe the demonstration of a prototypeDNN-based indoor localization system for floor-level location estimation usingreal received signal strength (RSS) data collected at one of the buildings onthe XJTLU campus. The preliminary results for both building/floorclassification and floor-level location estimation clearly show the strengthsof DNN-based approaches, which can provide near state-of-the-art performancewith less parameter tuning and higher scalability.