Recognition of Outlying Driving Behaviors: A Data-Driven Perspective with Applications to V2X Collective Perception - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Recognition of Outlying Driving Behaviors: A Data-Driven Perspective with Applications to V2X Collective Perception

(1) , (2) , (3)
1
2
3

Résumé

With Collective Perception, intelligent vehicles and road infrastructure can share and receive sensor data from their road peers to enhance mutual awareness. As traffic situations are incredibly dynamic and because of the potentially large number of objects on the road, the V2X channel can become quickly congested if vehicles choose to broadcast frequently. Even though many redundancy mitigation and congestion control techniques have been proposed, another thought is to only broadcast relevant safety-critical information about traffic situations that nearby drivers should be aware of. In this paper, we attempt to recognize outlying driving behaviors that imperil the safety of nearby road participants. The method is data-driven, does not require hard-coding of traffic rules, thus can be quickly adapted to different markets and scenarios while minimizing the incidence of delay until detection.
Fichier non déposé

Dates et versions

hal-03727824 , version 1 (19-07-2022)

Identifiants

Citer

Thinh Hoang Dinh, Vincent Martinez, Daniel Delahaye. Recognition of Outlying Driving Behaviors: A Data-Driven Perspective with Applications to V2X Collective Perception. 2021 IEEE Vehicular Networking Conference (VNC), Nov 2021, Ulm, Germany. pp.52-59, ⟨10.1109/VNC52810.2021.9644627⟩. ⟨hal-03727824⟩

Collections

ENAC OPTIM ANITI
15 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook Twitter LinkedIn More