K-means clustering for mitigation of nonlinear phase noise in digital coherent optical systems using 16-QAM modulation format

This work analyzes the use of the K-means clustering algorithm to mitigate nonlinear phase noise in single-span coherent systems, such as long-reach passive optical networks (LR-PONs). Simulations revealed that for a 100-km LRPON employing 16-ary quadrature amplitude modulation (QAM) and considering a 1:64 splitting ratio, the adoption of K-means with K-means++ initialization achieves an optimum bit error ratio (BER) of 6.310-4, whereas employing maximum likelihood, 10-3 is obtained. We also show that in order to achieve this performance improvement in 90% of the cases, K-means requires only 2,000 symbols.