K-means-clustering-based fiber nonlinearity equalization techniques for 64-QAM coherent optical communication system

Area of Science:

• Optical communication systems engineering
• Signal processing algorithms
• Nonlinear optics

Background:

• Fiber nonlinearity significantly degrades signal quality in high-capacity optical networks.
• Mitigating nonlinear impairments is crucial for achieving higher data rates and longer transmission distances.
• Existing methods often face challenges with complexity and optimization.

Purpose of the Study:

• To propose novel k-means clustering algorithms for mitigating fiber nonlinearity in 64-quadrature amplitude modulation (64-QAM) signals.
• To experimentally validate the effectiveness of these algorithms in a 75-Gb/s coherent optical communication system.
• To assess the performance improvements in terms of bit-error-ratio (BER) under various conditions.

Main Methods:

• Development of two k-means-clustering-based algorithms: training-sequence assisted and blind k-means.
• Experimental demonstration in a 75-Gb/s 64-QAM coherent optical communication system over 50-km single-mode fiber.
• Measurement of BER performance with varying launched powers to evaluate nonlinearity mitigation.

Main Results:

• The proposed k-means algorithms demonstrated reduced clustering complexity and data redundancy.
• Algorithms efficiently found optimal centroids, achieving global solutions for large k values.
• Significant mitigation of amplified spontaneous emission noise and fiber Kerr nonlinearity was observed, leading to improved BER performance.

Conclusions:

• K-means clustering provides an effective approach for fiber nonlinearity mitigation in 64-QAM signals.
• The developed algorithms offer a practical solution for enhancing the performance of coherent optical communication systems.
• These techniques contribute to more robust and efficient optical data transmission.