Disturbance/Uncertainty Observer-based Control of Vehicle Yaw Dynamics with Limited External Moment

This study addresses a novel strategy to enhance the accuracy and reliability of direct yaw-moment controller for vehicle directional stability. In the proposed method, some complementary terms are added to the bicycle model, as a reduced-order vehicle model to upgrade its content toward the actual model. The controller designed by the enhanced bicycle model compensates for perturbations such as unmodelled dynamics, parametric uncertainties and wind forces. Therefore, a reliable model-based controller is achieved to control vehicle directional stability in various maneuvers. The limitations of external yaw moment are modeled using a constrained optimization problem. Simulation studies are conducted in the CarSim software considering detailed vehicle dynamics and wind forces. The results demonstrate the superior efficiency of the constrained controller, developed by the enhanced bicycle model, compared with the researches in which the perturbation is not estimated for being used by the controller. During a severe lane-change maneuver, for instance, estimating the disparities between the bicycle model and the CarSim model causes the root mean square of the yaw rate tracking error to decrease from 0.237 to 0.005 )rad(. Importantly, the calculated external yaw moment  is also reduced from 980.13 to 557.4 )N.m). In another path tracking test, a comparison with the linear quadratic method, indicates 74% reduction in the RMS of the yaw rate tracking error and  19% reduction in the RMS of the control effort. The obtained results imply that the proposed controller can remarkably improve the tracking accuracy with much lower control efforts through estimating the model perturbations.