Robust Longitudinal Control Design Using Dynamic Inversion and Quantitative Feedback Theory

Abstract

The objective is to design low-bandwidth control laws that yield robust performance over a wide e ight envelope in the presence of 20% model uncertainties. This is achieved by using a hybrid controller in which a secondorder dynamic inversion transforms the nonlinear angle-of-attack dynamics to linear time-invariant form, while an outer loop, designed using quantitative feedback theory (QFT), ensures robust performance. The dynamic inversion reduces the plant variations due to nonlinearity, which are traditionally treated as uncertainty in QFT. Thereduced uncertainty allows lowercontrollerbandwidth fora givenlevel of performance. Thehybrid controller design is applied to the unstable, nonlinear, short-period dynamics of the F-16 aircraft and compared with a baseline QFT controller designed without dynamic inversion. Both designs provide satisfactory performance, but the baseline QFT controller requiresa bandwidth of around 4 ‐8 rad/s, whereasthe hybrid design has a bandwidth of only about 1 rad/s, leading to reduced demands on the control hardware. The success of the dynamic inversion depends strongly on the desired linear dynamics, and so recommendations are given on their selection. Nomenclature Cmtotal = nondimensional pitch moment coefe cient Cm®;Cmq;Cm±e = nondimensional partial derivative of Cmtotal with respect to ®;q, and ±e CXtotal;CZtotal = nondimensional coefe cient of x=z force N c = mean aerodynamic chord Iyy = moment of inertia about pitch axis M = aerodynamic pitch moment, M D N qSN cCmtotal m