/dnn-based_indoor_localization

Research on deep-neural-network-based indoor localization

Primary LanguageJupyter NotebookMIT LicenseMIT

DNN-Based Indoor Localization

This repository is for the implementation of advanced DNN models for scalable indoor localization with Wi-Fi fingerprinting. In this work, we are to exploit single-input and multi-output models1 for flexible control of loss weights for different outputs; they are natural extensions of the multi-lable classifier that was investigated in the previous work.

Footnotes

1 A special case of multi-input and multi-output models.