This article proposes a novel receding-horizon reinforcement learning approach for time-delayed human–machine shared control of intelligent vehicles. First, we build a multikernel-based data-driven ...