A Time-Varying Driving Style Oriented Model Predictive Control for Smoothing Mixed Traffic Flow

Connected and automated vehicles (CAVs) have great potential to smooth mixed traffic flow. To focus on the fact that the driving styles of HDVs are time-varying, a new control framework based on time-varying model predictive control (MPC) is proposed for mixed traffic flow in longitudinal control. Firstly, calibrating the parameters of car-following model based on genetic algorithm with the offline trajectory data as HDVs’ long-term driving style. Secondly, the short-term driving style is recognized by modifying the long-term driving style based on the real-time trajectory data. Then, a time-varying MPC control framework was proposed for smoothing mixed traffic flow. Finally, compared to the baseline model, the method we proposed achieves high performance both in fuel consumption and driving comfort in an emergence braking scenario. Numerical results show that the fuel consumption reduced by over 11.28% in a large-scale traffic system with a 17%–33% penetration rate of CAVs based on our method.