Research Article
Analyzing Driving Behavior: Towards Dynamic Driver Profiling
@INPROCEEDINGS{10.1007/978-3-030-67369-7_13, author={Anas Ouardini and Imane El Ouazzany Ech-chahedy and Afaf Bouhoute and Ismail Berrada and Mohamed El Kamili}, title={Analyzing Driving Behavior: Towards Dynamic Driver Profiling}, proceedings={Ad Hoc Networks. 12th EAI International Conference, ADHOCNETS 2020, Paris, France, November 17, 2020, Proceedings}, proceedings_a={ADHOCNETS}, year={2021}, month={1}, keywords={Spherical KMeans clustering Driver profiling Local Dynamic Map}, doi={10.1007/978-3-030-67369-7_13} }- Anas Ouardini
Imane El Ouazzany Ech-chahedy
Afaf Bouhoute
Ismail Berrada
Mohamed El Kamili
Year: 2021
Analyzing Driving Behavior: Towards Dynamic Driver Profiling
ADHOCNETS
Springer
DOI: 10.1007/978-3-030-67369-7_13
Abstract
This paper aims to use driving data to create a profile of the driver behavior, which can be then added as an additional layer to the Local Dynamic Map of the vehicle. The main contribution of the paper consists of using theSphericalKMeansClustering, an unsupervised clustering algorithm for multidimensional datasets, to segment the continuous driving data into multiple segments (hyperspheres). Unlike the state of the art, this helps in studying the behavior since all the data will be processed at the same time regardless of the number of features. The generated hyperspheres are an abstract form of the initial numerical values, and can be contribute to a better representation of the driver behavior. We used the UAH Dataset [9] to present the proposed approach, and the cross-validation technique to evaluate the segmentation results.
