Quaternion-based adaptive output feedback attitude control of spacecraft using Chebyshev neural networks

Area of Science:

• Aerospace Engineering
• Control Systems
• Artificial Intelligence

Background:

• Spacecraft attitude control is crucial for mission success.
• Uncertainties and external disturbances pose significant challenges.
• Existing adaptive control methods may suffer from over-estimation problems.

Purpose of the Study:

• To develop robust adaptive output feedback controllers for uncertain spacecraft attitude control.
• To enhance control performance using Chebyshev neural networks (CNN).
• To address the over-estimation issue present in some adaptive neural network controllers.

Main Methods:

• Utilizing quaternion representation for singularity-free global attitude description.
• Employing a nonlinear reduced-order observer to estimate output derivatives.
• Implementing two adaptive neural network controllers (CNN-based) with smooth robust compensators.
• Developing a standard adaptive controller for comparative analysis.

Main Results:

• Both CNN-based controllers ensure uniform ultimate boundedness of all closed-loop system signals.
• Adaptive NN controller-II effectively mitigates the over-estimation problem of controller-I.
• Simulations demonstrate superior performance of the proposed CNN-based approach over standard adaptive methods.

Conclusions:

• Chebyshev neural networks offer a robust and effective solution for uncertain spacecraft attitude control.
• The proposed output feedback controllers provide enhanced stability and performance.
• The adaptive NN controller-II presents an improvement by avoiding over-estimation issues.