We present the use of a neural-network-based controller for the F8 aircraft, with special emphasis on piloted Eight. Linearized longitudinal dynamics of the F8 aircraft has been used for this study, and the primary objective is to obtain a decoupled response to pilot pitch rate and velocity commands. An exact inverse neurocontroller with full state feedback was synthesized and was found to satisfy all control design objectives. Its ability to learn the inverse dynamics during the course of flight, greatly increases its survivability, The controller is found to provide stable and near desired response in the presence of modeling uncertainty less or equal to 30, actuators sluggishness, and battle damaged or partially missing Eight control surfaces. References: 28
展开▼
