Personalised lane keeping assist strategy: adaptation to driving style - Université Polytechnique des Hauts-de-France Accéder directement au contenu
Article Dans Une Revue IET Control Theory and Applications Année : 2019

Personalised lane keeping assist strategy: adaptation to driving style

Résumé

The research in advanced driver assistance systems (ADAS) has progressed to design architectures for driver specific performance. Personalised ADASs in this aspect have been developed with adaptation to driver attributes, state, style, behaviour, skill and so on. For the lane keeping task, the driver driving style while navigating a high/low curvature track plays an important part in the design of a lane keeping assist system. Considering this aspect, a robust co-operative control approach is formulated to design a personalised lane keeping assist with adaptation to driver style. Based on statistical analysis of lateral jerk and steer feel, a fuzzy rule based identification procedure for the classification of the driver style as clam, moderate, aggressive or very aggressive is designed. Using the identified driving style, a modulation function is proposed to adapt the assist torque. The assist torque is generated based on a robust higher order sliding mode approach as a feedback control for the driver-vehicle system. Closed-loop stability of the proposed driver-vehicle design in the presence of disturbances is established. Co-operative control between human driver-autonomous controller for the lane keeping task over the Satory test track with adaption to driver style is then shown for validation of the proposed architecture.
Fichier non déposé

Dates et versions

hal-03644349 , version 1 (19-04-2022)

Identifiants

Citer

Jagat Jyoti Rath, Chouki Sentouh, Jean-Christophe Popieul. Personalised lane keeping assist strategy: adaptation to driving style. IET Control Theory and Applications, 2019, 13 (1), pp.106-115. ⟨10.1049/iet-cta.2018.5941⟩. ⟨hal-03644349⟩
9 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More