Making Sense of the Noise - Integrating Multiple Analyses for Stop and Trip Classification
Published in Free and Open Source Software for Geospatial , 2022
Spang, R. P., Pieper, K., Oesterle, B., Brauer, M., Haeger, C., Mümken, S., Gellert, P. & Voigt-Antons, J.-N.
We propose a robust pipeline for classifying stops and trips from noisy real-world GPS and accelerometer data. By combining density-based clustering, temporal smoothing, and sensor-fusion heuristics, the approach improves segmentation stability across heterogeneous devices and sampling rates. Open-source tools and benchmarks on daily-life datasets demonstrate higher precision/recall and practical defaults for mobility research.
Recommended citation: Spang, R. P., Pieper, K., Oesterle, B., Brauer, M., Haeger, C., Mümken, S., Gellert, P. & Voigt-Antons, J.-N. (2022, August). Making Sense of the Noise - Integrating Multiple Analyses for Stop and Trip Classification. Paper presented at the Free and Open Source Software for Geospatial (FOSS4G 2022). Firenze, Italy. https://doi.org/10.5194/isprs-archives-XLVIII-4-W1-2022-435-2022