A Data-based Lane-keeping Steering Control for Autonomous Vehicles: A Human-in-the-Loop Approach

In this paper, we study the design of a lane-keeping control assistance system, by taking the interaction between the driver and the vehicle into account. A Human-in-the-loop controller design approach is proposed to assist the driver to maintain in the central position of a lane. A sampled-data adaptive dynamic programming (ADP) method is introduced to develop online adaptive optimal controllers to achieve zero steady-state lane-keeping error while the road curvature condition and the vehicle dynamics are unknown. The effectiveness of this steering control assistance system is supported by rigorous analysis and validated by simulation results. Furthermore, as opposed to previous Human-in-the-loop controllers, the proposed controller is capable of adapting to different road curvature conditions.