An Anti-Disturbance Adaptive Control Approach for Automated Vehicles in Mixed Connected Traffic Environment

In the foreseeable future, a coexistence of human-driven vehicles (HVs) and connected automated vehicles (CAVs) is expected in traffic flow systems. Effectively controlling CAVs to improve overall traffic performance and stability is a crucial yet challenging issue, especially when considering the uncertain behavior of HVs. This study proposes an optimal CAV controller for mixed connected traffic based on the adaptive model predictive control (AMPC) method. To mitigate the instability of the mixed platoon, the concept of loose disturbance string stability (LDSS) is introduced and integrated into the controller. Instead of using a fixed platoon size, a platoon formation criterion is established to dynamically calculate the number of vehicles in a platoon, taking into account the uncertain behavior of HVs, such as sudden deceleration and lane changes. An adaptive parameter tuning approach is incorporated into the MPC-based control framework to enhance the control performance in terms of stability, accuracy and disturbance recovery ability. The effects of the proposed CAV controller on overall traffic performance are investigated through two simulation experiments. In the first experiment, the performance of the proposed controller and LDSS are verified under three scenarios: sudden deceleration, vehicle leaving, and vehicle cut-in. In the second experiment, the proposed AMPC method is compared with two existing models: the linear quadratic regulation -based and acceleration-based connected cruise control models. The comparison results demonstrate that the proposed control model ensures LDSS and outperforms the other two approaches in terms of disturbance mitigation, albeit with a slight sacrifice of traffic efficiency.