Physics-driven untrained neural network for vortex beam compensation in adaptive optics aided underwater wireless...

Area of Science:

• Optical Communications
• Adaptive Optics
• Turbulence Compensation

Background:

• Orbital angular momentum (OAM) multiplexing is key for high data capacity in underwater wireless optical communications (UWOC).
• Underwater turbulence causes wavefront distortions, degrading OAM mode orthogonality and communication performance.
• Existing adaptive optics methods often require extensive datasets or iterative processing.

Purpose of the Study:

• To introduce a physics-driven, untrained learning approach for adaptive optics in UWOC.
• To enable accurate retrieval of underwater turbulence characteristics without pre-trained datasets or iterative processing.
• To validate wavefront recovery technology for compensating OAM beams under various turbulence strengths.

Main Methods:

• A physics-driven untrained learning approach for adaptive optics.
• Utilizing a one-shot distorted probe beam and its a priori known amplitude.
• Employing a hybrid input/output alternating projection algorithm with a square constraint area for phase screen retrieval.

Main Results:

• Accurate retrieval of underwater turbulence characteristics from a single distorted probe beam.
• Successful wavefront recovery and phase screen generation beyond the [0, 2π] range, validated across diverse turbulence strengths.
• Demonstrated compensation of OAM beams through numerical simulations and optical experiments, with inferred BER and channel capacity.

Conclusions:

• The proposed adaptive optics method effectively compensates for underwater turbulence in OAM-based UWOC systems.
• This approach offers a robust and efficient solution for improving the reliability and capacity of underwater optical communication.
• The validated wavefront recovery technology paves the way for practical implementation of high-performance UWOC.