A neural network composite learning control method for hypersonic aircraft based on high gain observer

Abstract

The invention discloses a high supersonic speed aircraft neural network composite learning control method based on a high gain observer and aims to solve a technical problem of poor practicality of ahigh supersonic speed aircraft control method in the prior art. The method is advantaged in that a strict feedback mode of an attitude subsystem is transformed to acquire an output feedback mode, a high gain observer is utilized to estimate unknown variables, and base is provided for subsequent controller design; system lumping uncertainty is considered, only one neural network is needed for approximation, the controller is simple to design, and engineering realization is facilitated; a system modeling error is introduced, composite new neural network learning rules are constructed, stable high supersonic speed aircraft neural control under uncertain situations is realized, and good practicality is realized.

Description

technical field

[0001] The invention relates to a hypersonic vehicle control method, in particular to a hypersonic vehicle neural network composite learning control method based on a high-gain observer. Background technique

[0002] As a high-precision weapon with rapid strike capability, hypersonic aircraft has attracted great attention from many military powers. Due to the integrated design of the engine / body, coupled with the complex dynamic model and flight environment, the hypersonic vehicle has the characteristics of strong nonlinearity and strong uncertainty. These characteristics make the design of hypersonic vehicle controllers face great challenges. Therefore, the handling of uncertainty is crucial to the safe flight of hypersonic vehicles.

[0003] As a typical control method, the backstepping method is widely used in the control of hypersonic vehicles. However, there are inherent defects in the traditional backstepping design. Using the backstepping method to...

Images