Citation

Sago T, Arai Y, Ueyama Y, Harada M. Trans. Soc. Automot. Eng. Jpn. 2023; 54(6): 1281-1286.

Copyright

(Copyright © 2023, Society of Automotive Engineers of Japan)

DOI

10.11351/jsaeronbun.54.1281

PMID

unavailable

Abstract

This paper investigates real-time optimal feedback control for an autonomous vehicle. The applicability of a deep learning neural network controller using the simplified model open-loop optimal control solution as supervised learning data is investigated for path following and obstacle avoidance maneuvers on the road, including straight and curved sections. The constructed controller can obtain optimal control variables for given states and constraints in real-time without iterative computation. The numerical results using the full vehicle model show that the proposed controller has the potential for real-time optimal obstacle avoidance control of the conventional type of vehicle.

Language: ja

Keywords

Full-Vehicle Simulation; Lane-Keeping Assistance System; Motion Control; Optimal Feedback Control; Vehicle Dynamics