Hierarchical Stage-Wise Training of Linked Deep Neural Networks for Multi-Building and Multi-Floor Indoor Localization Based on Wi-Fi RSSI Fingerprinting

In this article, we present a new solution to the problem of large-scale multibuilding and multifloor indoor localization based on linked neural networks, where each neural network is dedicated to a subproblem and trained under a hierarchical stagewise training (HST) framework. When the measured data from sensors have a hierarchical representation as in multibuilding and multifloor indoor localization, it is important to exploit the hierarchical nature in data processing to provide a scalable solution. In this regard, the hierarchical stagewise training framework extends the original stagewise training framework to the case of multiple linked networks by training a lower hierarchy network based on the prior knowledge gained from the training of higher hierarchy networks. The experimental results, with the publicly available UJIIndoorLoc multibuilding and multifloor Wi-Fi received signal strength indicator (RSSI) fingerprint database, demonstrate that the linked neural networks trained, under the proposed hierarchical stagewise training framework, can achieve a 3-D localization error of 7.98 m, which, to the best of the authors’ knowledge, is the most accurate result ever obtained for neural network-based models trained and evaluated with the full datasets of the UJIIndoorLoc database, and that, when applied to a model based on hierarchical convolutional neural networks, the proposed training framework can also significantly reduce the 3-D localization error from 11.78 to 8.71 m. The generalization capability of the proposed framework, for different localization scenarios, is also demonstrated with the UTSIndoorLoc single-building and multifloor database.