engleski

Spatial Machine Learning Personal Mobility Predictive Model Trained with Smartphone-Collected Trajectory Data

Individual and group mobility is an essential information for numerous segments of technology (including transport and logistics), society, and economy. The ability of telecommunications devices, such as smartphones, to collect accurate and reliable data on personal mobility with the embedded sensors, inspires research in personal mobility. We confirm the ability of suitably defined indicators to compare sets of trajectories, and identify outliers/differences among the individual ones. Furthermore, we demonstrate development of a machine learning (ML) regression predictive model based on experimental data collected on the real urban environment of the city of Krapina, Croatia, suitable for utilisation in personal mobility analysis, and traffic and transport planning and optimisation.

GNSS, trajectory, kinematics, machine learning, predictive model, mobility

The manuscript summarises the Erasmus+ project results, accomplished by student Mr Boyan Stoyanovski, while on his Erasmus+ traineeship at Krapina University of Applied Sciences, Krapina, Croatia, in Summer 2022.