Kick-scooters detection in sensor-based transportation mode classification methods
Abstract
In this work we present a novel classification model that can detect kick-scooters from inertial and pressure sensors. The detection is performed with kick-scooters being trained with other activities and transportation modes including still, walking, biking, taking bus and tramway. Results show that kick-scooters can be precisely detected up to 99% for three different sensor placements: on-foot, waist-attached and in the trouser's pocket. Thus, this paper provides a first contribution where kick-scooters can be classified and studied for further applications such as mobility behavior analysis and navigation.
Origin | Files produced by the author(s) |
---|