Real-time deployable and robust cooperative control strategy for a platoon of connected and autonomous vehicles by factoring uncertain vehicle dynamics

Abstract

In real-world driving environments, a platoon of connected and autonomous vehicles (CAVs) will be subject to many disturbances (e.g., aerodynamic drag, dynamic road gradients) that can prevent them from achieving desired control accurately. These disturbances can cause uncertainties in the vehicle dynamics, reducing platoon safety and mobility. To address this problem, this study proposes a robust cooperative control (RCC) strategy, developed as a minimax problem, to ensure the safe and efficient maneuvering of a CAV platoon in the worst-case situation due to uncertainties in the vehicle dynamics. The maximization subproblem of the minimax problem determines the inputs of the uncertainties to maximize the cost function (i.e., the platoon performance indicator), while the minimization subproblem seeks to optimize the control decisions for all following vehicles in the platoon to minimize the cost function using inputs of the uncertainties from the maximization subproblem. A novel partition-based method of feasible direction is proposed to solve the minimax problem. It is globally convergent and computationally efficient, enabling the RCC strategy to be deployed in real time. The conditions for robust stability of the CAV platoon are also analyzed. Results from numerical studies show that compared to a control strategy that ignores uncertainties in vehicle dynamics, the RCC strategy can substantially improve the platoon performance under disturbances. Hence, it can be used to enable the safe and efficient maneuvering of CAV platoons in a real-world driving environment.