Cellular Echoes: How Your Phone Navigates Without GPS express

Original: "Cellular Signal Constructed Convolutional Vision Transformer for High Accuracy Positioning"
arXiv:2607.12518 · 2026-07-14 · CC BY-SA 4.0 · 1 min
A neural echolocator determines location from signal reflections with an accuracy of up to 3 meters.

📄 Showing the "Simple" version — "Popular" is not ready yet. Add it to favorites to help prioritize it.

A city signal bounces between walls like a ball, and your phone catches dozens of its 'reflections'—echoes. Each microsecond of delay adds 300 meters to the path.

The signal's speed is constant, like that of light (speed of light). Knowing the delay and angle, you can reconstruct the pattern of reflections.

The neural network processes this noise in two stages. First, like a keen ear, it filters out interference and isolates each echo. Then it assembles them into a coherent picture, much like delay analysis splits a radio signal into 'rays'.

Measuring power at different delays resembles brightness estimation (brightness estimation), while comparing signal paths is like detecting invisible objects by flickering (transit method). This is how a bat's brain constructs a 3D image of space.

City tests: distance error—3.5 m, direction error—2.5°. That's enough to tell which side of a building you're on. Signals from multiple towers are combined and smoothed for better accuracy. A regular smartphone will soon navigate indoors as if it had GPS.

Every microsecond of delay is 300 meters, so by listening to echoes, your phone acts like a radar with centimeter-level precision.

Philip K. Dick foresaw this technology: in his stories, police tracked people by signals — now this becomes reality without any satellites.

Scientists
Christian DopplerD. B. McLaughlinDidier QuelozMichel MayorR. A. RossiterJames Clerk Maxwell
Concepts
Convolutional Neural Networks cellular neural interface transformers wavelet transform retina support vector machine antennas very long baseline interferometry