We propose a blind joint equalization algorithm for M-QAM signals based on a widely linear filtering approach. The proposed scheme jointly compensates receiver IQ imbalance and polarization mixing, along with carrier recovery, followed by transmitter IQ imbalance compensation. We first investigate the proposed scheme's tolerance to transceiver IQ Imbalance, polarization mixing, phase noise and frequency offset through numerical simulations for 32 GBd PM-16QAM and PM-64QAM signals and compare its performance with the conventional digital processing algorithms. Further, with the proposed algorithm, we experimentally demonstrate the improvement in Q<sup>2</sup> value to up to ~ 1.22 dB for a 32 GBd PM-16QAM and ~ 3.72 dB for a 16 GBd PM-64QAM signal with a phase imbalance of 9<sup>o</sup>. We show that the MSE convergence of the proposed joint equalizer is much faster than conventional DSP algorithms. Deployment of such an equalizer in optical communication systems is beneficial due to its improved tolerance to multiple impairments, albeit with increased complexity.
Trellis-based feed-forward carrier recovery for coherent optical systems
