Learning-enhanced hybrid control for beam jitter suppression in LEO-GEO optical communication links

Area of Science:

• Spacecraft engineering
• Control systems engineering
• Optical communication

Background:

• Micro-vibrations cause beam jitter in optical communication links, degrading performance.
• Accurate disturbance identification is crucial for effective vibration suppression.
• Existing methods may struggle with non-stationary disturbances and computational load.

Purpose of the Study:

• To propose a physics-guided hybrid control framework for real-time suppression of micro-vibration-induced beam jitter.
• To enhance control performance and stability in long-distance optical communication links.
• To minimize computational overhead while maintaining high precision.

Main Methods:

• A disturbance-observer-based Linear Quadratic Integral architecture with output-only autoregressive identification.
• Integration of a lightweight neural module (custom Transformer) as a bounded residual compensator.
• Implementation of a gating mechanism for stable fallback operation under non-stationary disturbances.

Main Results:

• Achieved high identification correlation (up to 0.9997) under multi-condition disturbances.
• Reduced Root Mean Square (RMS) jitter from 0.324918 to 0.04841 µrad.
• Introduced only 0.3% additional computational overhead compared to conventional methods.

Conclusions:

• The proposed framework effectively suppresses micro-vibration-induced beam jitter in optical communication links.
• The hybrid approach offers superior performance and efficiency for spaceborne optical systems.
• Results demonstrate the potential for reliable, high-precision beam stabilization in demanding environments.