Gain scheduling for lateral motion of propulsion controlled aircraft using neural networks

A neural network approach to gain-scheduling linear dynamic controllers for the lateral motion of propulsion controlled aircraft (PCA) is introduced. The PCA system is adopted for emergency flight control of airplane with multiple control surface failure in lateral motion. The linear controllers are synthesized at distinct flight conditions by the H_∞ methodology which is applied to the problem of matching the dynamically compensated throttle-actuated crippled aircraft and the nominal control-surface-actuated aircraft model. The various H_∞ controllers at various flight conditions are used to train a radial basis network (RBN) which is then used as gain scheduling controller. A simulation on the lateral control design of an L-1011 under emergency fly-by-throttle control demonstrates the concept