Modeling and On-line Solution of Air Combat Optimization Problems and Games

The thesis deals with the guidance and control of a fighter aircraft in air combat. From the modeling perspective, the thesis formulates realistic air combat optimization problems and games in which inherent uncertainties of air combat are taken into account. From the solution perspective, the thesis develops on-line methods for the near-optimal feedback solution of the constructed models. The solution methods are based mostly on receding horizon control, where computational savings are achieved by using a truncated planning horizon. Considering singlesided optimization problems, the thesis presents a new approach for the guidance of an aircraft avoiding a homing air combat missile. The approach is applicable with various avoidance criteria that exploit different weaknesses of the missile system. In addition, a novel way based on Bayesian reasoning for taking into account the target’s uncertainty about the guidance law of the missile is introduced. An approach and its software implementation for the user-oriented computation of realistic near-optimal aircraft trajectories are presented as well. The software can be used to assess the quality and realism of the solutions provided by the introduced guidance schemes. Considering games, the thesis presents an influence diagram game modeling a dogfight between two aircraft and develops a method for the on-line optimization of the aircraft’s controls in such a setting. The game formulation enables the consideration of preferences, perception, and beliefs in air combat. In addition, a new game model and an on-line solution scheme providing game optimal support time of the missile in a missile duel is introduced. Considering practical implementation, the introduced models and solution methods could be further developed and consolidated in an onboard guidance system or in a pilot advisory system.