My final-year project: an Android app that navigates inside large concrete buildings where GPS simply fails — using nothing but the Wi-Fi already in the building.
How it works
Maps Wi-Fi signal strengths across the building as location "fingerprints".
Uses a K-Nearest-Neighbour algorithm to estimate the user's position.
Reaches room-level accuracy across multiple floors — no extra hardware needed.
Impact
Graded A- for technical complexity, with successful real-world testing throughout the InfoLab21 building.
THE CHALLENGE
Draft: confirm details — GPS dies indoors. Students and visitors got lost in the multi-floor InfoLab21 buildings, and beacon-based navigation systems cost more than the university wanted to spend.
THE TECHNICAL SOLUTION
Draft: confirm details — built a mobile app that locates users by fingerprinting the existing Wi-Fi infrastructure and classifying position with machine learning — no beacons, no new hardware.
KEY BUSINESS OUTCOME
The app located users to the correct room across the multi-floor InfoLab21 buildings in real-world testing — using only the existing Wi-Fi infrastructure, no beacons or extra hardware.
