Trial Input Method and Own-Aircraft State Prediction in Autonomous Air Combat

The history of the trial maneuver method in autonomous air-combat technique is traced. Development of an adaptive maneuvering logic program and tactical guidance research and evaluation system is introduced. Several immanent deficiencies of the trial maneuver method are discussed in the paper. To avoid these deficiencies, an improvement of a new trial input method with trial control-surface deflection input is presented in detail in this paper. This paper also introduces a neural network system that predicts own-aircraft state based on measurements of present state and the trial control-surface deflection input. Simulation results show that this prediction framework performs well; particularly, this system is able to predict movement of the aircraft with large sideslip angles, which the trial maneuver method has failed to accomplish.