Optimal control based on artificial neural network for flameout approach ff fighter jet

Abstract

Loss of thrust due to engine failure induces dangerous situation for a single-engine fighter jet. For this reason, particularly almost every military aircraft has well established flameout approach procedures destined to the base or an alternate airfield. The flameout approach procedures enables the aircraft to safely return to the destination while ensuring utmost safety. The control strategy consisting the procedures is rather a recommendation provided to a pilot, rather than an optimal control strategy. Hence, detailed control techniques are different depending on pilots. Because the optimization process to calculate an optimal control strategy is accompanied by a massive computation time, it is almost impossible to calculate and apply the optimal control onboard in real-time. In this thesis, an optimal control strategy based on real-time optimal control method for conducting the flameout approach procedures under arbitrary flight situation is proposed by using the artificial neural network which is actively studied as a part of the artificial intelligence research.