Authors:
Shangyuan Zhang
1
;
2
;
Makhlouf Hadji
1
and
Abdel Lisser
2
Affiliations:
1
Institut de Recherche Technologique SystemX, 8 Avenue de la Vauve, 91120 Palaiseau, France
;
2
CentraleSupelec, L2S, Université Paris Saclay, 3 Rue Curie Joliot, 91190, Gif-sur-Yvette, France
Keyword(s):
Adaptive Cruise Control, Distributionally Robust Optimization, Stochastic Optimization, Autonomous Vehicle.
Abstract:
Due to the recent advances in intelligent and connected vehicles, Adaptive Cruise Control (ACC) has become a key functionality of advanced driver-assistant systems (ADAS) to enhance comfort and safety. The evaluation of ACC’s efficiency and safety is also crucial for the industry to prove the reliability of its products. In our paper, we propose a distributional robust optimization-based ACC reference generation model to produce the optimal commands facing the uncertainty of sensors. By taking into account the uncertainty set with knowledge of the first and second moments, the original optimization problem with chance constraints can be simplified and solved more efficiently. Numerical experiments in a driving simulator illustrate that the robustness of the results is largely increased by minimizing the risks of violation of safety constraints.
