The 20th and 21st joint COTA International Conference of Transportation Professionals (CICTP 2020-21)

Full connected and automated vehicle (CAV) market penetration rate can be highly unrealistic in the near future which leads to a transition period of mixed traffic environment. Thereby, this paper proposes a decentralized model predictive control (DMPC) strategy for heterogeneous vehicle platoon pertaining of CAVs and human driven vehicles (HV) under a dynamic two-predecessor-following information topology (DTIT) considering communication failure. The illustrious intelligent driver model (IDM) is applied to model and predict HV’s driving behavior. Three various scenarios of platoon composition are developed depending on the different percentage and positions of HVs. The objective cost function is designed by penalizing on the deviations between the actual and desired trajectories and constraints are proposed to help stabilize the whole platoon system and attenuate oscillation more effectively. The application of DTIT depicts the flexibility of system communication and the effectiveness of control degradation avoidance. Several field test based numerical simulated experiments demonstrate the developed DMPC outperforms than the normal cooperative adaptive cruise control (CACC) strategy without prediction.

CICTP