A performance-based control design approach for nonlinear systems with safety-driven probabilistic requirements is proposed in this paper. Unlike conventional design methods where uncertainties are conservatively handled, the extensive probabilistic nature of uncertainties is propagated effectively using a higher-order unscented transform and the Pearson system. With the designed approach, the restrictions of Gaussian assumption is surpassed, and more accurate approximations are achieved with less computational effort. Closed loop statistical performance is guaranteed while maximizing the safety margin through optimization. Simulations are carried out on a longitudinal dynamic model considering parametric uncertainties to fully demonstrate the superior capability of ensuring closed loop statistical performance.     «

A performance-based control design approach for nonlinear systems with safety-driven probabilistic requirements is proposed in this paper. Unlike conventional design methods where uncertainties are conservatively handled, the extensive probabilistic nature of uncertainties is propagated effectively using a higher-order unscented transform and the Pearson system. With the designed approach, the restrictions of Gaussian assumption is surpassed, and more accurate approximations are achieved with le...     »