Conflict resolution for connected automated vehicles at unsignalized roundabouts considering personalized driving behaviours

Abstract

To address the driving conflicts of connected automated vehicles (CAVs) at unsignalized roundabouts, a cooperative decision-making framework is proposed. The personalized driving preferences of CAVs are considered in the decision-making algorithm, which are reflected by different driving styles. A motion prediction algorithm is used to improve the decision-making performance. The effect of the motion prediction algorithm on the decision-making performance is evaluated, including the advancement of driving safety and the computational load for the hardware. The cooperative game theoretic approach is applied to the interaction modelling and collaborative decision making of CAVs. Finally, hardware-in-the-loop (HIL) tests are carried out to evaluate the feasibility and real-time performance of the proposed algorithm.