Data-driven control for nonlinear automated vehicles with multi-description coding for handling data dropouts

This paper proposes a data-driven control (DDC) strategy for nonlinear automated vehicles, employing a multidescription coding (MDC) mechanism based on scalar quantization to address the challenges of data dropouts and limited bandwidth in networked communication environments. The developed MDC-based communication protocol enhances system robustness by reducing the probability of data dropout. It achieves this by transmitting multiple descriptions of source data through diverse channels, incorporating quantization and index reassignment to efficiently alleviate bandwidth constraints. Based on the reconstructed real-time data, a novel data-driven controller and a parameter estimation algorithm are designed, offering adaptability to varying driving conditions and rapid response in dynamic environments. A numerical simulation showcases the potential of the proposed approach as a reliable and efficient solution for the speed problem in nonlinear automated vehicles with challenging communication environments.