Deep learning-enabled atmospheric turbulence compensation for concentric perfect optical vortex beams

Area of Science:

• Optics and Photonics
• Optical Communications
• Atmospheric Optics

Background:

• Concentric perfect optical vortex (CPOV) beams are crucial for optical communications.
• Atmospheric turbulence (AT) severely distorts CPOV beam propagation, degrading free-space optical (FSO) communication performance.
• Effective mitigation of AT impacts on CPOV beams is a persistent challenge.

Purpose of the Study:

• To propose and validate a novel network for predicting atmospheric turbulence phase screens.
• To mitigate the detrimental effects of atmospheric turbulence on CPOV beams in FSO systems.
• To enhance the reliability and accuracy of OAM-based FSO communication systems.

Main Methods:

• Development of an atmospheric turbulence decomposition frequency prediction network (ATDFPNet).
• Training the ATDFPNet on an extensive dataset for accurate turbulence phase screen prediction.
• Integration of ResNet18 with a Dammann vortex grating for orbital angular momentum (OAM) mode identification.

Main Results:

• The ATDFPNet demonstrated rapid and precise prediction of turbulence phase screens with robust generalization.
• Mode purity of the CPOV outer ring improved from 8.00% to 38.52% under strong turbulence after AT compensation.
• OAM mode identification accuracy increased from 47.77% to 98.01%, and a communication link achieved a Structural Similarity index of 0.98.

Conclusions:

• The ATDFPNet effectively mitigates atmospheric turbulence distortions in CPOV beams.
• The proposed method shows significant potential for enhancing OAM-based FSO communication systems.
• This work advances the practical application of CPOV beams in challenging atmospheric conditions.