Evaluating Alternative Location Tracking Methods by Comparing GPS Data with Mobile Data

This thesis presents a comparative evaluation of GPS-based and mobile-based tracking methods across three core dimensions: accuracy, cost, and usability. The methodology used both an older dataset from 2017–2020 and a controlled experiment conducted in 2025 across seven road segments in Sweden. It appliedspatial distance errors, root mean square error (RMSE), mean absolute percentage error (MAPE), and sampling frequency evaluation. The results showed that GPS achieved notable positional accuracy in the older rural datasets (Mean Distance Error = 2.51 m for GPS vs. 3.27 m for mobile), aligning with previous studies. However, mobile-based tracking outperformed GPS in the new rural dataset ((Mean Distance Error = 3.68 m for GPS vs. 2.72 m for mobile), likely due to enhanced 4G/5G coverage and advancements in smartphone hardware and software. While GPS demonstrated stable and consistent temporal frequency (mean = 16 s for GPS vs. 23 s for mobile), both methods exhibited notable altitude errors (RMSE = 14.45 m for GPS vs. 29.09 m for mobile). Both tracking methods displayed similar speed trends (MAPE = 17.72% for GPS vs. 20.25% for mobile), though GPS performed slightly better than mobile in high-speed zones. Mobile tracking proved to be user-friendly and cost-effective, but it required proper setup, and in some cases continuous monitoring to ensure reliable and complete data. The study concludes that each tracking method offers specific strengths depending on the application context. GPS performed better in areas with limited infrastructure, while mobile tracking showed stronger performance in regions with good network coverage. Combining both tracking methods may enhance data reliability across different conditions.