Researchers at the University of Surrey have developed an artificial intelligence system capable of determining precise locations in dense urban areas without relying on GPS, reducing localization errors from 734 meters to within 22 meters. The innovation could support autonomous vehicles, delivery robots, and other navigation technologies.
The system, called PEnG (Pose-Enhanced Geo-Localization), combines satellite and street-level imagery to determine location using only visual data, according to a paper published in IEEE Robotics and Automation Letters. PEnG is designed for environments where GPS signals are weak or obstructed, such as tunnels, densely built cities, or areas with limited connectivity.
“Many navigation systems depend on GPS, but coverage isn’t always guaranteed. Our goal was to develop a solution that works reliably using only visual information,” said Tavis Shore, a postgraduate researcher in AI and computer vision at the University of Surrey. “By combining satellite and ground-level imagery, PEnG achieves a level of accuracy previously thought unachievable without GPS—and could help unlock new possibilities for autonomous vehicles and smart navigation tools.”
Unlike previous approaches limited by the frequency of satellite image updates, PEnG uses a two-step process, first narrowing down location at street-level and then refining it using relative pose estimation, a technique that analyzes the camera’s position and orientation. The system works with standard monocular cameras commonly found in vehicles.
“One of the most exciting aspects of this system is how it turns a simple monocular camera into a powerful navigation tool,” said Dr. Simon Hadfield, associate professor and primary supervisor of the project. “PEnG is designed to operate without GPS, making it ideal for fast-moving, unpredictable scenarios. That kind of flexibility is exactly what’s needed for the next generation of autonomous vehicles and robotics operating in challenging environments.”
The University of Surrey said the research has been released as open source to encourage further innovation in navigation technologies. Professor Adrian Hilton, director of Surrey’s Center for Vision, Speech and Signal Processing, added, “The ability to accurately pinpoint a location without GPS lays the foundation for smarter, more resilient autonomous systems that can operate in even the most remote environments.”
Shore and his team are now working on a functional prototype, supported by the university’s Ph.D. Foundership Award, which funds early-stage development of the GPS-free navigation device.
Resource:
- Tavis Shore et al, PEnG: Pose-Enhanced Geo-Localisation, IEEE Robotics and Automation Letters (2025). DOI: 10.1109/LRA.2025.3546513
