Design and experimental validation of adaptive iterative learning control for nonlinear vehicular platoons

This paper focuses on addressing repetitive vehicular platoon control problems for connected and automated vehicles (CAVs). Inspired by the repeat-and-refine mechanism in natural systems, we propose an adaptive iterative learning control (AILC) approach for CAVs to mainly overcome time-varying uncertainties and communication delays of nonlinear platoons. A novel adaptive updating law, combining the time-domain and iteration-domain adaptations, is established to both provide a reliable warm start for the initial iteration and enable error convergence across iterations. Vehicle stability is validated through the Lyapunov-like analysis, where a new candidate function is designed to explore the boundedness property. Furthermore, initial platoon errors are considered and compensated by designing a reference platoon, and the learning-based string stability is guaranteed by the proposed learning algorithm. Simulations demonstrate the efficiency of the algorithms adopted, and hardware experiments with electric wheeled vehicles further justify the practicability of the AILC method.