We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
An intelligent divide-and-conquer approach for driving style management.
- Authors
Al Abri, Khalid Ali; Jabeur, Nafaa; Gharrad, Hana; Yasar, Ansar Ul-Haque
- Abstract
Driving styles reflect the performance and the ability of drivers to drive in a safe and protective manner. As some of them would possibly result into harmful behaviors, the recognition of these styles continue to attract intensive investigations from the transportation community. In spite of the current promising results, the existing approaches did not yet address the management of simultaneous driving behaviors that are exhibited by a driver during the same commute. They did not also explicitly investigate the legal implication of these driving styles. To this end, we argue that intelligent collaborative solutions could adequately handle the constantly changing traffic environment, prevent aberrant driving behaviors, classify driving styles, and identify the right road traffic policies to apply at the right time to the right driver. Therefore, we are proposing a new intelligent divide-and-conquer approach that aims to process concurrent driver's driving behaviors and identify the related driving styles, accordingly. Our solution relies on a four-layer Multi-Agent System (MAS) architecture, where intelligent agents execute injection, filtering, action, and feedback processing steps to ultimately generate personalized recommendations and feedback to drivers. For the sake of illustration, we collected driving data about braking and acceleration behaviors via our dedicated mobile app AWARIDE. We successfully classified the driving styles into aggressive, normal, and conservative. We also successfully identified the transitions between these styles.
- Publication
Personal & Ubiquitous Computing, 2023, Vol 27, Issue 5, p1729
- ISSN
1617-4909
- Publication type
Article
- DOI
10.1007/s00779-023-01740-1