Optimal Flight Control with Nonlinear Actuator Compensation through Model Reference Adaptive Control

摘要

An approach for flight control system design in the presence of undesirable actuator nonlinearities is presented and demonstrated through simulation. A Robust Servomechanism Linear Quadratic Regulator (RSLQR) optimal control approach is applied to design a pitch-rate command tracking control system for a nonlinear F-16 aircraft model. A second-order actuator model with a deadband is implemented, and the impacts on actuator and aircraft longitudinal dynamic responses are shown to be degraded, with undesirable persistent oscillations (limit cycling). A model reference adaptive controller (MRAC) using a single layer neural network (SL NN) is applied for low-level actuator control, and simulation results show the RSLQR flight control system augmented with MRAC actuator control achieves desirable responses.