Influence of PID controlled CAV platoon on mixed traffic flow

With the rapid advancement of automated driving technologies, mixed traffic flow has increasingly become the new norm in transportation systems. Connected Autonomous Vehicles (CAVs), leveraging their connectivity and cooperative perception capabilities, demonstrate substantial potential in enhancing traffic efficiency and safety. This study develops a co-simulation platform based on CARLA and SUMO, where CAVs are modeled using the Intelligent Driver Model (IDM), Human-Driven Vehicles (HDVs) are reproduced through a driver-in-the-loop approach, and CAV platoons are controlled via position-based Proportional–Integral–Derivative (PID) control. Furthermore, a genetic algorithm-based vehicle positioning optimization method is applied on two-lane and three-lane road networks to construct simulation scenarios covering different CAV penetration rates (0.1, 0.3, 0.5, 0.9) and traffic saturation levels (0-1). The results indicate that under identical saturation conditions, increasing CAV penetration contributes to overall speed improvement; however, when the penetration rate reaches 0.5 at a saturation level of 0.6, the average speed of all vehicles decreases simultaneously, revealing a bottleneck effect induced by interaction complexity. In addition, HDVs exhibit greater sensitivity to saturation changes, whereas CAV platoons controlled by the PID demonstrate superior stability and efficiency. These findings can inform traffic management, optimize CAV deployment, and guide infrastructure planning to enhance efficiency and safety under varying penetration rates and traffic saturation.