While GPS is the de facto standard outdoor localization method, we are still lacking a similar indoor technology. In this project, we aim to close this gap. We use signal-processing, machine learning and deep-learning based algorithms to build a ubiquitous calibration-free indoor localization system. We aim to identify the user floor-level and her 2D location within that floor.
As part of the project, we proposed a pseudo-3D WiFi-based indoor localization. For deployment, the system only requires the building’s floorplans and WiFi APs locations. The system was tested in (up to 9 floors) 5 buildings in a period spanning 6 months by 13 subjects. It achieved significant improvements over state-of-the-art systems. The system combines a deep neural network and an RSS-Rank Gaussian-based method to estimate the user location.
Publications:
“Hapi: A Robust Pseudo-3D Calibration-Free WiFi-based Indoor Localization system”
Heba Aly and Ashok Agrawala
EAI MobiQuitous, 2018
“TrueStory: Accurate and Robust RF-based Floor Estimation for Challenging Indoor Environments”
Rizanne Elbakly, Heba Aly and Moustafa Youssef
IEEE Sensors Journal, 2018Journal
“An Analysis of Device-Free and Device-Based WiFi-Localization Systems”
Heba Aly and Moustafa Youssef
International Journal on Ambient Intelligence and Computing, 2014 InvitedJournal
“New Insights Into Wifi-based Device-Free Localization”
Heba Aly and Moustafa Youssef
CoSDEO workshop. In ACM, UbiComp Adjunct., 2013
“Demo: New DfP localization insights”
Heba Aly and Moustafa Youssef
CoSDEO workshop. In ACM, UbiComp Adjunct., 2013
- Acceptance Rate: 20%
- Best Paper Runner Up