A data-driven switching control approach for braking systems with constraints?

Braking control is of paramount importance in guaranteeing driving safety and comfort, but it is a well-known challenging task, due to the highly nonlinear and road conditiondependent behavior of the vehicle. Existing braking controllers typically rely on accurate models of the vehicle dynamics and the vehicle-road interaction, which are quite difficult to be retrieved in practice. In the wake of the data-driven control paradigm, we propose a model-free and fully data-based braking control method. The architecture of our scheme is two-layered, featuring: an inner switching controller, directly designed from data to match a given closed-loop behavior, and an outer predictive reference governor, exploited to enforce constraints and possibly improve the overall braking performance. The effectiveness of the approach is shown in a simulation environment, by providing a sensitivity analysis to the main tuning knobs of the method. (c) 2022 Elsevier Ltd. All rights reserved.