Interactive control strategy of mixed traffic with connected and automated vehicles at unsignalized intersections: a double-layer framework

The absence of traffic signals at unsignalized intersections often results in operational conflicts that reduce traffic efficiency. The advent of connected and automated vehicles (CAVs) offers new opportunities for managing such intersections. However, because human-driven vehicles (HVs) still dominate traffic flow, the coexistence of CAVs and HVs poses significant challenges for mixed traffic control. This paper proposes a double-layer framework for interactive control of mixed CAV-HV traffic flows at unsignalized intersections. At the upper layer, a centralized control module employs a topological sorting algorithm to assign vehicle passage priorities and prevent deadlock situations. At the lower layer, a distributed control module formulates a decision-making model that balances efficiency and safety while capturing interactions among CAVs, HVs, and between CAVs and HVs. Based on this framework, vehicle trajectories and conflict points within the intersection are computed to derive dynamically optimal control actions. Simulation results demonstrate that the proposed control strategy ensures safe passage and improves intersection efficiency as CAV penetration rates increase. Compared with conventional HV traffic, a fully CAV traffic flow under the proposed strategy increases average speed by 1.16 times and throughput by 1.24 times. Furthermore, under growing traffic demand, the strategy consistently enhances mixed traffic efficiency at unsignalized intersections, with particularly strong benefits under high-demand conditions.