Sound Plus Radio Could Finally Make Indoor GPS Accurate

GPS works great outdoors but turns useless inside buildings, where signals weaken and bounce off walls. To solve this, engineers have looked for alternatives like Wi-Fi, but those are still imprecise. A team from Belgium and Sweden built a new tool to help: ARFT, a dataset that records radio and ultrasound signals at the same time, in the same room, across thousands of spots. Think of it as a giant training ground for indoor positioning systems.

In their test room, the ceiling was covered with 42 antennas and 91 microphones. A small speaker, serving as a rover, moved to over 5,000 positions. At each stop, it sent out an ultrasonic chirp while radio signals were recorded from every antenna. Because all measurements are perfectly synchronized, researchers can teach AI to locate something using sound alone, radio alone, or both together. The radio part works like advanced Wi-Fi sensing, and the sound part works like sonar — but miniaturized and indoors.

The researchers also showed a basic method to calculate position from the sound data: catch the chirp with multiple microphones, measure tiny arrival time differences, then solve a math puzzle to get coordinates. This is just a starting point — the main value is in the shared dataset, which lets other teams build smarter positioning systems without spending months setting up their own equipment.

Why should you care? Indoor navigation still relies on dumb visual cues or beacons. With better training data, your phone could guide you through a huge mall to the exact store, warehouse robots could find shelves without maps, and first responders could locate someone in a smoky building. The catch? This is bleeding-edge research data, not a product in your pocket. But it's a meaningful step toward a world where GPS stops working only at the front door.