The Quest for Sub-Meter: Benchmarking High-Precision GNSS Modules for Use with Firefighters

Problem Statement: 

Commercial smartphones have a localization error of ~5 meters, which is insufficient for precise localisation of firefighters during missions. The market is currently flooded with low-cost, high-precision COTS (Commercial Off-The-Shelf) GNSS modules claiming centimetre-level accuracy using RTK and PPP technologies. However, manufacturer datasheets are often overly optimistic, and most of these claims are unrealistic or only achievable in optimal conditions.

Tasks

Research GNSS error sources, RTK/PPP correction methods, and benchmarking metrics for localisation systems.

Set up a testing system on a Raspberry Pi platform.

Conduct rigorous, real-world testing of the most promising modules (e.g., u-blox ZED-F9P vs. Unicore UM980) in urban environments.

Work Distribution

20% Theory: Research on GNSS error sources, RTK/PPP correction methods, and benchmarking metrics for localisation systems.

70% Prototyping & Testing: Set up a multi-module testing rig on a Raspberry Pi, develop scripts for automated NMEA data collection, and conduct real-world tests.

10% Writing: Compare performance against ground-truth data and provide a verdict on what module is the best.

This assignment is set in collaboration with Inertia Technology B.V.

Contact:

Berend Jan van der Zwaag (b.j.vanderzwaag@utwente.nl)

Sabari Anbalagan (sabari@inertia-technology.com)