Enhancement of Nonlinear Smooth Trackers Using Neural Network for Post LOC Autonomous Recovery of Flying Vehicles

摘要

The stance of modern aeronautical and aerospace vehicles require advanced level of automation to increase their levels of maneuverability beyond human capabilities, ensure their safety and mitigate upset flight regimes where crew skills may not be appropriated. Inappropriate not because their lack of training, but because their flying vehicles are going through flight regimes with heavy nonlinearities and standard procedures become obsolete.Those heavy nonlinearities induced flying vehicles towards upset flight regimes where Loss - Of - Control(LOC) is likely to occur due unstable zero dynamic/sliding dynamic often difficult to be fully captured during the design with classical affine models and flight control techniques. As a consequence, we isolate off - line upset flight regimes, design and embed intelligent flight control laws that would ensure that flying vehicles regain control by tracking a particular safe trajectory or regular trim within the safe set of the flight envelope.The longitudinal dynamic of the Generic Transport Model(GTM) model is used for illustration of post LOC autonomous flight recovery regimes.