Russian Scientists Build a GPS-Free Transit Tracking System Using Computer Vision
A team in Perm has created an AI-powered monitoring platform that identifies city buses in real time — no satellites required.
Researchers at Perm Polytechnic University have unveiled an alternative to GPS for tracking public transport. Their new system uses computer vision and AI algorithms to recognize buses through existing road and city surveillance cameras, transmitting live updates to passengers and dispatchers — even in areas with weak or no GPS signal.
The innovation tackles a common pain point in transit systems worldwide: unreliable GPS data caused by connectivity issues or power failures. Instead of depending on satellite tracking, the Russian-developed solution uses visual recognition, analyzing camera footage to determine the exact position and route of a vehicle.
AI That Sees Through the Storm
The system identifies bus and trolley numbers with 82% accuracy, even in low visibility conditions such as rain, snow, or nighttime lighting.
To reduce errors, the AI doesn’t rely on a single frame — it confirms each bus number multiple times before sending data. All information is instantly relayed via a chatbot interface, giving commuters accurate arrival times.
Because the system works with existing traffic and security cameras, it’s far cheaper to implement than GPS networks. It runs on standard office computers, using less than 10% of CPU resources, making it easily scalable for cities worldwide.
In a country known for long distances and extreme weather, this project from Perm may point toward a new generation of low-cost, camera-based smart mobility systems — no satellite required.
