Efficient Real-Time Road Curvature Estimation : Visual-Inertial Approach - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Efficient Real-Time Road Curvature Estimation : Visual-Inertial Approach

Résumé

Road curvature is an essential parameter of road geometry, and it is crucial to set vehicle design and operating speeds. It can be utilized to identify the maximum stable velocity of a Powered Two Wheeled Vehicle (P2WV) and predict other safety-related events, including Lane departure and lane crossing. This paper proposes a new vision-based approach to estimate the road’s curvature accurately and efficiently under real-time constraints for P2WV. The proposed method is based on the vanishing point approach to estimate the relative heading and its dynamics. Combined with the available vehicle speed and the yaw rate given by the inertial measurement unit (IMU), the instantaneous curvature of the road is reconstructed. The proposed algorithm is then tested using various simulated scenarios of different speeds and curvatures to validate the approach. Then it was compared to other estimation methods based on Inverse Perspective Mapping (IPM) to investigate the validity and efficiency in all scenarios regarding accuracy and time complexity. The proposed method shows very promising results in terms of error and real-time execution.
Fichier principal
Vignette du fichier
IFACWorldCongress2023_OA_AC_LN_HA_Efficient_RealTime_Road_Curvature_Estimation _VisualInertial_Approach.pdf (2 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04068430 , version 1 (13-04-2023)

Identifiants

Citer

Obaida Alrazouk, Amine Chellali, Lamri Nehaoua, Hichem Arioui. Efficient Real-Time Road Curvature Estimation : Visual-Inertial Approach. The 22nd World Congress of the International Federation of Automatic Control (IFAC 2023), Jul 2023, Yokohama, Japan. pp.4953--4958, ⟨10.1016/j.ifacol.2023.10.1270⟩. ⟨hal-04068430⟩
145 Consultations
159 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More