Loading...
Adaptive neural network control of discrete-time nonlinear systems
ZHANG JIN
ZHANG JIN
Citations
Altmetric:
Alternative Title
Abstract
In this thesis, adaptive neural network control schemes are investigated for five classes of discrete-time nonlinear systems in affine/non-affine form. The systems studied include single-input single-output nonlinear systems, multi-input multi-output nonlinear systems. For affine systems, the existence of the desired feedback controls is proved. For non-affine systems, by using implicit function theorem, the existence of the implicit desired feedback controls is also proved. Single layer neural networks, such as radial basis function neural networks, high order neural networks, and multi layer neural networks are used as the emulators for the desired feedback controls. State feedback, output feedback, backstepping design are used respectively in developing the stable adaptive controls. By using Lyapunov method, semi globally uniformly ultimate boundedness stability is proved for each kind of systems. All the closed loop signals are guaranteed to be bounded. Simulation results, including numerical simulation for two practical affine/non-affine continuous stirred tank reactors systems, show the effectiveness of the proposed control schemes.
Keywords
adaptive control, neural networks, discrete-time systems, SISO systems, MIMO systems, SGUUB
Source Title
Publisher
Series/Report No.
Collections
Files
Loading...
ZhangJin-Thesis.pdf
Adobe PDF, 1.64 MB
Rights
Date
2004-08-24
DOI
Type
Thesis