Flight Control Using Physical Dynamic Inversion

Abstract

This paper presents a new architecture of dynamic inversion based flight control law, which decouples desired dynamics specification and stabilizing feedback controller design. The decoupling is achieved by using a physically integrated reference model, which mimics desired aircraft response, and an incrementally cascaded error controller, which shapes the extra dynamics beyond the desired dynamics. This approach makes best use of flight dynamics both in feed forward and feedback design, in an attempt to provide design transparency and to ease the feedback control efforts, and thus named Physical Dynamic Inversion here. Pilots commands are filtered out by the explicit reference model when getting close to the envelope boundaries, together with reducing feedback commands at the same time, flight envelope protection are completed. The novel features are demonstrated by comparison with the existing architectures in the literature in the aspects of desired dynamic specification, loop closure and ideal closed loop dynamics. Rules of gain tuning are also introduced, the benefits of auto gain scheduling of dynamic inversion based control law preserves, while providing satisfactory flying qualities. Nomenclature A = Aerodynamic frame B = Body-fixed frame O = NED frame E = Earth centered earth fixed frame K = Kinematic frame ̅ = Rotated kinematic frame = Flight path angle = Flight path bank angle = Pitch angle = Elevator deflection = Damping ratio = Aerodynamic speed of the Center of gravity = Kinematic speed of the Center of gravity ̅ = Modified Flight path angle rate, ̅ =̇/cos ( ) = Pseudo Control Subscripts abs = Absolute act = Actuator c = Command signal des = Desired ff = Feed forward signal n = Nominal r = Reference signal sp = Short period extra = Extra command compared to signal limit alw = Allowed magnitude of the signal 1 Ph.D Candidate, Institute of Flight System Dynamics(FSD), Email: fubiao.zhang@tum.de, Student Member AIAA. 2 Professor, Institute of Flight System Dynamics(FSD), Email: florian.holzapfel@tum.de, Senior Member AIAA. Downloaded by CORNELL UNIVERSITY on July 31, 2015 | http://arc.aiaa.org |

Adaptive Neural Network‐Based Satellite Attitude Control by Using the Dynamic Inversion Technique and a VSCMG Pyramidal Cluster

“…The role of this robustness component is to limit the values of the neural network's matrices and . Equation (59) has been deduced by Calise et al [29,30,33] for applications involving adaptive control laws based on dynamic inversion technique and neural networks.…”