PAM-4 Eye-Opening Monitoring Techniques Using Gaussian Mixture Model

To achieve next-generation high-speed data transmission standards, e.g., IEEE 802.3bs and PCIe6.0, four-level pulse amplitude modulation (PAM-4) data formats are adopted. Although PAM-4 signaling is spectrally efficient in mitigating the inter-symbol interference (ISI) caused by the bandwidth-limited wire channels, PAM-4 is more sensitive compared with conventional non-return-to-zero (NRZ) binary signaling. In this paper, to evaluate the received signal quality for adaptive coefficients setting of an equalizer for PAM-4 data transmission, a novel eye-opening monitoring (EOM) technique based on machine learning is proposed. The monitoring technique uses a Gaussian mixture model (GMM) to classify the received PAM-4 symbols. Moreover, simulation and experimental results of the coefficients adjustment using the method are presented.