A Comparison of Adaptive Nonlinear Control Designs for an Over-Actuated Fighter Aircraft Model

Abstract

In this paper three nonlinear adaptive flight control laws based on two design philosophies and incorporating various types of control allocation algorithms are designed for a simplified nonlinear over-actuated fighter aircraft model. The control designs are Lyapunov based to ensure closed-loop stability, boundedness of system states and tracking error convergence. The designs differ in the way the on-board aircraft model is updated to changes due to e.g. control surface failures or aircraft damage. The first two controllers are based on an integrated adaptive design; i.e. the identifier is designed simultaneously with the control law using a control Lyapunov function and is thus driven by the tracking error. The third control law design is based on another design philosophy: the identifier and the control law are designed as separate modules. Identification for this modular design is done using a recursive least squares method. Nonlinear damping terms are used in this control framework to compensate for varying parameter estimates, since the certainty equivalence principle does not hold for nonlinear systems. Second-order actuator dynamics are taken into account in the control designs. The control designs are evaluated using numerical simulations where several cases of locked control surface failures are considered during two different maneuvers. No sensor information about these failures is fed back to the control system. A comparison of tracking performance and especially estimation accuracy of the design methodologies is made. Three different types of control allocation algorithms, based on either the pseudo-inverse or quadratic programming, are incorporated in each of the adaptive control designs during the simulations to study their impact on system performance.