Global, robust and intelligent control method used for elastic aircraft and based on singular perturbation strategy

Abstract

The invention relates to a global, robust and intelligent control method used for an elastic aircraft and based on a singular perturbation strategy. The method analyzes an aircraft elastic body dynamical model based on a singular perturbation theory, performs decoupling on rigid and flexible modes and realizes fast and low time scale separation. Specific to an elasticity fast subsystem, a slidingmode controller is designed; specific to an attitude slow subsystem, neural network control in an effective approximation region and robust control outside the effective approximation region are designed based on a switching mechanism, and global stability of a closed-loop system is realized; meanwhile, based on a system tracking error and a predication error, weight of a neural network is updated, and learning performance of the neural network is improved.

Description

technical field [0001] The invention relates to an aircraft control method, in particular to an elastic aircraft global robust intelligent control method based on a singular perturbation strategy, and belongs to the field of aircraft control. Background technique [0002] With the development of lightweight aircraft materials, the problem of structural elasticity of the system is becoming more and more obvious. Structural elasticity will not only cause the sensitive components of the system to receive elastic information, resulting in a decrease in the control accuracy of the system, but more seriously, it may cause the fuselage to break. Therefore, it is necessary to Elastic modes are effectively handled. Neural networks can accurately approximate unknown nonlinearities and are widely used to solve aircraft control problems. However, most control methods based on neural networks can only achieve semi-global stability of the closed-loop system, and cannot guarantee the contr...

AI Technical Summary

Problems solved by technology

[0002] With the development of lightweight aircraft materials, the problem of structural elasticity of the system is becoming more and more obvious. Structural elasticity will not only cause the sensitive components of the system to receive elastic information, resulting in a decrease in the control accuracy of the system, but more seriously, it may cause the fuselage to break. Therefore, it is necessary to Efficient handling of elastic modes

Neural networks can accurately approximate unknown nonlinearities and are widely used to solve aircraft control problems. However, most control methods based on neural networks can only achieve semi-global stability of closed-loop systems, and cannot guarantee the control performance of the system.

"Robust adaptive neural control of flexible hypersonic flight vehicle with dead-zone input nonlinearity" (Bin Xu, "Nonlinear Dynamics", (2015) 80:1509-1520) designed adaptive neural network control for the longitudinal channel model of elastic hypersonic vehicle, Treat the elastic mode as a disturbance item, and use robust adaptive estimation to deal with the influence of the elastic mode, but this processing method lacks in-depth analysis and research on the dynamics of elastic bodies, and cannot accurately deal with the elastic mode, and the The method does not consider whether the neural network approximation is valid, and cannot achieve an accurate approximation of unknown dynamics

Images