Accurate separation of mixed high-dimension optical-chaotic signals using optical reservoir computing based on...

Area of Science:

• Optoelectronics
• Nonlinear Dynamics
• Information Theory

Background:

• Optical chaotic signals are complex and challenging to separate.
• Optically pumped spin-Vertical Cavity Surface Emitting Lasers (spin-VCSELs) offer unique properties for generating chaotic signals.
• Reservoir computing (RC) is a powerful paradigm for processing complex time-series data.

Purpose of the Study:

• To propose and demonstrate schemes for separating mixed optical chaotic signals.
• To investigate the effectiveness of VCSEL-based RC systems for signal separation.
• To address scenarios with known and unknown mixing fractions.

Main Methods:

• Utilizing two parallel reservoirs based on spin-VCSELs with optical feedback and injection.
• Employing two cascaded RC systems for scenarios with unknown mixing fractions.
• Demonstrating separation of signals mixed with up to three beams of chaotic X-polarization components (X-PCs) and Y-polarization components (Y-PCs).

Main Results:

• Effective separation of two groups of mixed optical chaos signals with training errors no more than 0.093 when mixing fractions are known.
• Accurate prediction of mixing fractions using two parallel reservoirs in a cascaded RC system.
• Successful separation of mixed optical chaos signals with unknown mixing fractions, also achieving errors below 0.093.
• Demonstrated effective separation for signals superimposed with more than three chaotic beams.

Conclusions:

• VCSEL-based RC systems provide an effective platform for separating mixed optical chaotic signals.
• The proposed methods are robust for both known and unknown mixing fractions.
• This work offers potential advancements for multi-channel chaotic cryptography communication, including multiple access and demultiplexing.