Polarization demultiplexing by recursive least squares constant modulus algorithm based on QR decomposition

Recursive least squares constant modulus algorithm based on QR decomposition (QR-RLS-CMA) is first proposed as the polarization demultiplexing method. We compare its performance with the stochastic gradient descent constant modulus algorithm (SGD-CMA) and the recursive least squares constant modulus algorithm (RLS-CMA) in a polarization-division-multiplexing system with coherent detection. It is demonstrated that QR-RLS-CMA is an efficient demultiplexing algorithm which can avoid the problem of step-length choice in SGD-CMA. Meanwhile, it also has better symbol error rate (SER) performance and more stable convergence property.

► For the first time, QR-RLS-CMA is used for polarization demultiplexing. ► Symbol error performance of QR-RLS-CMA is superior to SGD-CMA and RLS-CMA. ► QR-RLS-CMA has better convergence property than SGD-CMA and RLS-CMA.