Minimal-learning-parameter technique based adaptive neural control of hypersonic flight dynamics without back-stepping

This paper investigates one robust adaptive controller for hypersonic flight dynamics. The altitude tracking is transformed into the control problem of the attitude subsystem, which is composed of flight path angle, pitch angle and pitch rate. Different from previous design using back-stepping related technique, this paper analyzed the tracking control without back-stepping where the controller is synthesized with high gain observer and minimal-learning-parameter technique. The highlight is that the design procedure is greatly simplified and the computation burden of parameter updating is reduced. It is proved that the filtered tracking error is guaranteed in the semiglobal sense. Simulation results are presented to demonstrate the effectiveness of the design.