Why GPS makes distances bigger than they are.

Area of Science:

• Movement analysis
• Geospatial data science
• Sensor technology

Background:

• Global Navigation Satellite Systems (GNSS), including the Global Positioning System (GPS), are crucial for tracking movement trajectories of vehicles, animals, and humans.
• GPS data is inherently susceptible to measurement and interpolation errors, impacting the accuracy of recorded movement parameters.
• High-frequency sampling in GPS data minimizes interpolation errors, making measurement errors more influential on distance calculations.

Purpose of the Study:

• To investigate the systematic bias in distance measurements derived from GPS data.
• To mathematically explain the phenomenon of distance overestimation in GPS trajectories.
• To introduce a novel method for quantifying the quality of GPS movement data based on error autocorrelation.

Main Methods:

• Mathematical modeling to explain the relationship between GPS measurement error and distance bias.
• Introduction of autocorrelation (C) as a key parameter for GPS error analysis.
• Development and application of a novel approach to determine C in real-world, high-frequency GPS data.

Main Results:

• GPS-recorded distances are systematically overestimated compared to true distances.
• The degree of overestimation is mathematically linked to the autocorrelation (C) of GPS measurement errors.
• Empirical analysis of pedestrian and car trajectories confirmed strong spatial and temporal autocorrelation in GPS measurement errors, providing a data quality estimate.

Conclusions:

• Measurement errors in absolute positioning systems, particularly GPS, introduce a systematic bias leading to distance overestimation.
• The autocorrelation of measurement error (C) serves as a valuable quality indicator for movement data.
• The findings are applicable beyond GPS to any absolute positioning system where interpolation errors are negligible.