Energy-saving optimization in mixed traffic flow using a dual-loop cascade control framework with predictive coordination

To optimize energy economy and control performance in light of the growing popularity of connected and autonomous vehicles (CAVs) in mixed traffic systems, we present a dual-loop cascade control architecture in this study (Quadratic variable exponential spacing strategy-based nonlinear model predictive control with distributed coupled adaptive fault tolerant control, QNDCC). The architecture combines distributed coupled adaptive fault-tolerant control (DCAFTC) and nonlinear model predictive control (NMPC). The inner-loop controller uses a DCAFTC strategy to handle the uncertainty brought on by sensor failures, while the outer-loop NMPC strategy integrates fuel consumption and car-following performance into the optimization framework. We build a car-following model taking multi-information uncertainty into account to accurately represent realistic elements like sensor failures, measurement noise, and actuator delays. Comprehensive simulation results under a variety of conditions, including emergency braking, temporary congestion, UDDS and HWFET drive cycles, and others, demonstrate that QNDCC greatly increases control reaction speed and tracking accuracy while lowering overall fuel consumption by up to 42.4 %. Furthermore, the strategy efficiently decreases the energy waste in the acceleration stage, achieves better energy distribution amongst queue, according to the analysis based on the energy consumption breakdown of vehicle operation phases. These findings demonstrate that QNDCC is a practical choice for intelligent transportation systems that require compatible energy consumption and car-following control.