Accurate positioning is easy to describe but surprisingly difficult to measure. A GNSS receiver may perform very well in open conditions while struggling in a city surrounded by tall buildings, beneath dense foliage or near multi-level roads where satellite signals are reflected and obstructed. For automotive applications, testing in these difficult environments is therefore just as important as measuring performance under ideal conditions.

FocalPoint's testing methodology for its Supercorrelation technology and S-GNSS Auto illustrates how complex this evaluation can become. A central challenge is establishing reliable ground truth: a reference position that is significantly more accurate than the receiver being tested. The company uses tightly coupled GNSS/INS reference systems combining high-grade inertial sensors with dual-antenna GNSS receivers and sophisticated post-processing.

Testing is conducted across different geographical and environmental conditions, including urban canyons, foliage, underpasses and open-sky routes. Multiple GNSS constellations and frequency bands are also considered, while the results are compared with commercial receivers from other manufacturers.

An important distinction is made between two levels of accuracy. Measurement-domain analysis examines raw GNSS observables such as pseudorange and Doppler before they are turned into a position. Position-domain analysis then evaluates the final position and velocity delivered to the user. Looking at both helps identify not only whether accuracy improves, but also why.

The analytical process combines statistical and causal methods. Cumulative distribution functions can show overall error behaviour at different percentiles, while detailed investigation of individual events can reveal rare failure modes hidden within otherwise strong results. Synthetic GNSS signals complement real-world testing by providing precisely known conditions for algorithm development and stress testing, although simulated results still need to be validated against live measurements.

This combination is particularly relevant as GNSS becomes an increasingly important input for ADAS, V2X and automated driving. Reliable positioning cannot be demonstrated through a single accuracy figure. It requires repeatable testing, credible ground truth and an understanding of what happens at the edges of system performance.
 
 
CROCKETT, Mark. How we test: FocalPoint's approach to rigorous GNSS performance evaluation. FocalPoint [online]. 2 September 2026 [cited 2026-09-18]. Available from: https://focalpointpositioning.com/resources/insights/blogs/how-we-test-focalpoints-approach-to-rigorous-gnss-performance-evaluation/

Picture: chat gpt