Design and Worst-case Validation of a Longitudinal Adaptive Flight Control Law : a practical approach

This paper describes the design and the worst-case validation of an adaptive longitudinal Flight Control Law (FCL) for a transport aircraft with multiple contraints: handling qualities, aeroelastic stability & structural loads issues, and limited computer workload. This adaptive controller covers the loss of flight dynamics information which are used to schedule the FCL gains. A new controller design is proposed, based on the LFT (Linear Fractional Transformation) framework. It allows a faster design while keeping the same controller structure, design methods and validation expertise as for previous aircraft. The LFT gains are combined with an improved Least-Square estimator of the aircraft model. These two elements provide adaptation to multiple sensor failures while demonstrating their robustness to realistic external disturbances. This adaptive scheme is not only tested around an aircraft equilibrium but also along trajectories. It is shown that the aircraft behaviour depends on the large set of possible initial conditions for the estimator. Thus a worst-case approach (based on genetic optimization algorithms) is considered to validate the transient and asymptotic aircraft nonlinear performance. All simulations are run on a nonlinear six degree-of-freedom certified Airbus simulator.