Navigation Without GPS That Already Works

Smartphones can determine indoor locations by scanning WiFi access points (using Google and Apple databases). But how do you calculate coordinates without GNSS, WiFi, or cellular signals? It turns out that in some cases, location can be determined using only accelerometer data.

Jamming and Spoofing

GPS has become unreliable. Jamming and spoofing in combat zones create interference "far beyond military conflict areas." Regional GPS interference maps show the scale of the problem as of April 2025.

Hobbyists with instructions on GitHub can spoof GPS signals for a couple hundred dollars. Governments actively distort navigation signals in electronic warfare. Spoofing has moved from academic papers into reality.

Alternatives to GPS

Alternative navigation methods are being developed to combat GPS signal suppression.

AQNav System: In 2024, SandboxAQ introduced the AQNav navigation system (developed over 1.5 years under a US Air Force contract). The system uses "AI algorithms, quantum sensors, and the Earth's crustal magnetic field to provide real-time geomagnetic navigation in the absence of GPS signal."

The key element is sensitive quantum magnetometers. AI algorithms match crustal patterns against known magnetic maps.

Quantum Accelerometers: The London Underground began testing the first quantum accelerometer in 2024. Inside a steel cabinet sits a container with several billion rubidium atoms and a laser cooling system.

Navigation by Stars

Engineers at the University of South Australia developed a celestial navigation system for unmanned aerial vehicles (published in the journal Drones on November 7, 2024). The prototype combines star triangulation with machine vision algorithms.

Advantages: Celestial navigation only works at night, but this is the preferred operating time for many UAVs. Unlike GNSS, the system "neither receives nor emits any signals, making it immune to existing jamming methods."

System Components:

• Raspberry Pi 5
• Alvium 1800 U-240 monochrome sensor
• 6mm f/1.4 wide-angle lens

Testing showed that the UAV could reliably determine its location with accuracy up to 4 km when flying at fixed altitude and speed. The system works under clear starry skies and can serve as a backup tool when interference occurs.

The Transit App

The Transit mobile app for finding public transportation accurately shows "the coordinates of subway trains and the user themselves by detecting vibrations." It requires no cellular connection, WiFi, or GPS.

How the Technology Works:

Accelerometer data (acceleration in m/s²) is converted to frequency in hertz via Fourier transform. When the train moves, the smartphone vibrates at a frequency of about 5 Hz; when walking, it's about 2 Hz.

Model Training: Two volunteers rode all the lines and stations of the New York subway over two weeks, collecting data from multiple iPhone and Android devices. A neural network was trained to distinguish types of vibration:

• Moving subway train
• Train during a stop
• Walking
• Walking up stairs
• Other types of movement

The Mixer Model: Takes four inputs into account:

1. Type of movement
2. Last known user location
3. How recent the last known location is
4. Train schedules

The model outputs the current location, "correctly predicting it in 90% of cases."

Both models run locally on the smartphone, require no network connection, and don't send data to the developer's servers. The system resembles inertial navigation in rockets and UAVs, where coordinates are calculated based on direction and speed without satellite signals.

Barometric Method: Several years ago, experiments were conducted with smartphone barometers. Thanks to the Venturi effect, these sensors "show with 100% accuracy the entry and exit of a train from a subway tunnel" even without accelerometers. Entry and exit correspond to stations, and the distances between them are known.

Limitations: This is a very limited form of "navigation without GPS" that only works in subway tunnels. However, it is one of the first examples of commercial deployment of such a system for real users, not just at the experimental stage.

Theoretically, the system could be expanded by training the model to account for surface transport, long-distance train schedules, typical user routes, and other factors.

A Curious Fact: GPS on the Moon

In March 2025, a GPS signal was successfully "captured" on the Moon for the first time as part of the LuGRE experiment on the Blue Ghost lander (Firefly Aerospace). LuGRE relies on GPS and Galileo satellite constellations for position triangulation.

The signal was detected at a maximum distance of 391,000 km. GNSS sensors will allow lunar vehicles to autonomously measure time, speed, and position without help from Earth-based operators — critical for a future lunar base.