Manoeuver planning, synchronized optimization and boundary motion control for autonomous vehicles under cut-in scenarios

This work proposes a two-phase motion control mechanism, which includes manoeuver preparation and maneuver execution phases. Compared with the state-of-art methodologies, this study has the following distinguishing features: Firstly, this method actively prepares a sequence of maneuvers ahead to be executed under cut-in scenarios with initially unsatisfied execution conditions. Therefore, the behavior-reachable region is considered as a constraint to resolve unsatisfied execution conditions with model predictive control in the maneuver preparation phase. Secondly, the proposed method designs the dynamic-related sliding boundary as the terminal region of model predictive control, enabling the system state-variables to enter the initial safety region of the barrier Lyapunov function. In this way, the barrier Lyapunov function-based controller can effectively finish driving tasks computationally while keeping state-variables constrained in a predefined zone in the next maneuver execution phase. Thirdly, with the designed time Synchronized Control Governor (SCG), the state-variables converge to the origin synchronically in finite time with less fluctuation/overshot. In the simulation, the parameter for dynamic-related sliding boundaries is first evaluated. The comparison between stabilization control and the time-synchronized control governor is made to verify superior performance of the proposed approach in fluctuation/overshot. Finally, three scenarios are simulated to be compared, which demonstrates the effectiveness of the designed dynamic-related sliding boundary.