Efficient PAM-4 Symbol Estimation Using Soft Clustering.

Multi-valued data transmission using four-level pulse amplitude modulation (PAM-4) is promising for alleviating the bandwidth limitation of interconnects. To evaluate the received signal quality during data transmission when using adaptive coefficient settings for a PAM-4 equalizer, we propose an eye-opening monitor method based on statistics of PAM-4 symbol transition. The proposed statistical eye-opening monitor is based on soft clustering and enables efficient PAM-4 data transmission under closed-eye conditions. Simulation results demonstrate the feasibility of symbol classification using the proposed statistical evaluation method.

Authors

• Yosuke Iijima
• Yasushi Yuminaka

Keywords

No keywords are indexed for this paper.