| Title |
Micromobility vehicle classification for smart city infrastructure using magnetic field sensor array |
| Authors |
Kasperavičius, Eidenis ; Navikas, Dangirutis ; Markevičius, Vytautas ; Valinevicius, Algimantas ; Žilys, Mindaugas ; Čepėnas, Mindaugas ; Ritonja, Jožef ; Klimenta, Dardan ; Hinov, Nikolay ; Andriukaitis, Darius |
| DOI |
10.1109/ACCESS.2026.3681387 |
| Full Text |
|
| Is Part of |
IEEE Access.. Piscataway, NJ : IEEE. 2026, vol. 14, p. 55641-55651.. ISSN 2169-3536 |
| Keywords [eng] |
machine learning ; magnetic sensors ; Micromobility ; smart cities ; traffic monitoring |
| Abstract [eng] |
Urban mobility worldwide is seeing an increasing use of micromobility vehicles such as bicycles and e-scooters. However, accurate methods for measuring their traffic parameters remain limited, which could result in outdated infrastructure planning practices. Conventional sensing technologies such as pneumatic tubes, computer vision, and inductive loops struggle with the small dimensions, low metal content, and high maneuverability of micromobility vehicles. This work presents an experimental classification method based on magnetic signatures, which is designed for micromobility traffic monitoring in smart city applications. A sensor array consisting of 11 magnetic field sensors installed across a bicycle path was used to collect 236 real-world vehicle signatures. Fourteen temporal, spatial, and frequency-domain features were extracted from each signature and evaluated using multiple machine learning models. Gradient Boosting model achieved the highest performance with a cross-validation accuracy of 82%, demonstrating that magnetic field signatures contain sufficient information to differentiate bicycles, e-scooters, and heavy micromobility vehicles. Even with a limited dataset, results demonstrate the feasibility of micromobility traffic monitoring based on magnetic signatures. Expanding the dataset and implementing additional signal processing algorithms is expected to further enhance classification accuracy and support safe and accessible mobility in smart cities. |
| Published |
Piscataway, NJ : IEEE |
| Type |
Journal article |
| Language |
English |
| Publication date |
2026 |
| CC license |
|
