Output-based enhanced closed-loop model reference adaptive control and its application to direct yaw moment control

The enhanced model reference adaptive control (EMRAC) algorithm is an effective full-state adaptive solution to control plants affected by nonlinear unmodelled dynamics and persistent disturbances. The EMRAC strategy improves the tracking performance by equipping the MRAC algorithm with adaptive switching and adaptive integral control actions. However, the need for the plant state prevents the applicability of the EMRAC algorithm to engineering control problems where only the plant output is measurable. To cover this gap and extend the range of plants controllable by EMRAC solutions, this article presents an output-based EMRAC algorithm leveraging the closed-loop (CL) MRAC formulation. The robustness of the closed-loop control system is analytically analysed, not only with respect to plant parameter uncertainties and square measurable disturbances, but also to & Laplacetrf;infinity$$ {\mathcal{L}}_{\infty } $$ unmodelled terms and disturbances. The ultimate boundedness of the closed-loop control system is assessed with respect to & Laplacetrf;infinity$$ {\mathcal{L}}_{\infty } $$ unknown nonlinear terms and disturbances, by using Lyapunov theory for Filippov systems, as the adaptive switching control action makes the closed-loop system discontinuous. To assess the effectiveness of the CL-EMRAC strategy to impose reference trajectories despite the unmodelled plant dynamics and persistent bounded disturbances, the problem of vehicles' direct yaw moment control is used as an engineering case study. The closed-loop tracking performance is also quantitatively evaluated through a set of key performance indicators and compared to those provided by four benchmark controllers, that is, two LQ-based strategies and two MRAC-based control solutions. The CL-EMRAC and benchmark controllers are implemented and tested in a co-simulation environment based on a high-fidelity IPG CarMaker vehicle model.