Smooth-Switching Control-Based Cooperative Adaptive Cruise Control by Considering Dynamic Information Flow Topology

Abstract

Vehicle-to-vehicle communications can be unreliable because of interference and information congestion, which leads to the dynamic information flow topology (IFT) in a platoon of connected and autonomous vehicles. Some existing studies adaptively switch the controller of cooperative adaptive cruise control (CACC) to optimize string stability when IFT varies. However, the difference of transient response between controllers can induce uncomfortable jerks at switching instances, significantly affecting riding comfort and jeopardizing vehicle powertrain. To improve riding comfort while maintaining string stability, the authors introduce a smooth-switching control-based CACC scheme with IFT optimization (CACC-SOIFT) by implementing a bi-layer optimization model and a Kalman predictor. The first optimization layer balances the probability of communication failure and control performance optimally, generating a robust IFT to reduce controller switching. The second optimization layer adjusts the controller parameters to minimize tracking error and the undesired jerk. Further, a Kalman predictor is applied to predict vehicle acceleration if communication failures occur. It is also used to estimate the states of preceding vehicles to suppress the measurement noise and the acceleration disturbance. The effectiveness of the proposed CACC-SOIFT is validated through numerical experiments based on NGSIM field data. Results indicate that the CACC-SOIFT framework can guarantee string stability and riding comfort in the environment of dynamic IFT.