Neural-Network-Based Adaptive Finite-Time Output Feedback Control for Spacecraft Attitude Tracking

Area of Science:

• Aerospace Engineering
• Control Systems
• Artificial Intelligence

Background:

• Spacecraft attitude control is critical for mission success.
• Challenges include actuator saturation, inertial uncertainty, and external disturbances.
• Existing methods may face computational complexity and filtering errors.

Purpose of the Study:

• To develop a neural network-based adaptive finite-time output feedback control for rigid spacecraft attitude tracking.
• To address actuator saturation, inertial uncertainty, and external disturbances.
• To ensure finite-time convergence of attitude tracking errors.

Main Methods:

• Design of a neural state observer to estimate unknown states.
• Application of adaptive neural finite-time command filtered backstepping (CFB).
• Use of compensation signals based on fractional power to mitigate filtering errors.

Main Results:

• Attitude tracking error converges to a desired neighborhood in finite time.
• All closed-loop system signals remain bounded in finite time, even with input saturation.
• Numerical simulations validate the algorithm's effectiveness.

Conclusions:

• The proposed neural network-based adaptive finite-time control is effective for spacecraft attitude tracking.
• The method successfully handles actuator saturation and uncertainties.
• The approach offers improved performance and stability guarantees.