Search

We proposed the ANN-based predistortion to mitigate nonlinear distortion in a 25- to 70-km IM/DD OFDM PON. Using 18-dBm launch power, the >55-Gbps downstream with 30-dB loss budget was demonstrated without nonlinear compensation at ONUs.

Volterra equalization (VE) presents substantial performance enhancement for high-speed optical signals but suffers from high computation complexity which limits its physical implementations. To address these limitations, we propose and experimentally demonstrate an elastic net regularization-based pruned Volterra equalization (ENPVE) to reduce the computation complexity while still maintain system performance. Our proposed scheme prunes redundant weight coefficients with a three-phase configuration. Firstly, we pre-train the VE with an adaptive EN-regularizer to identify significant weights. Next, we prune the insignificant weights away. Finally, we retrain the equalizer by fine-tuning the remaining weight coefficients. Our proposed ENPVE achieves superior performance with reduced computation complexity. Compared with conventional VE and L1 regularization-based Volterra equalizer (L1VE), our approach show a complexity reduction of 97.4% and 20.2%, respectively, for an O-band 80-Gbps PAM4 signal at a received optical power of −4 dBm after 40 km SMF transmission.

We employ ℓ0-regularization to reduce Volterra filter complexity by up to 90% in 112-Gbps PAM-4 VCSEL transmission. Compared to ℓ1-regularization, ℓ0-regularization achieves lower complexity and more precise weights without retraining after sparse identification.

A pruned Volterra equalizer that reduces computational complexity by up to 85.2% in 112-Gbps PAM-4 optical interconnects is proposed. The BER performance is competitive with LASSO-based Volterra equalizers, and training time is twice as fast.

A 400-Gbit/s Ethernet (GbE) pulse-amplitude-modulation-4 (PAM-4) over 10-km single-mode-fiber (SMF) transmission direct-detected by an integrated silicon-photonics-receiver satisfying the IEEE P802.3bs requirement; and its feasibility to support 512-Gbit/s were illustrated.

We have designed a novel convolutional neural network based nonlinear classifier that outperforms traditional Volterra nonlinear equalizers. A BER of 3.50 × 10-6 is obtained for a 112-Gbps PAM4 EML-based optical link over 40-km SMF transmission.

A noise-subspace-based adaptive beamformer is modified by introducing the triangular systolic array (TSA). A set of TSAs is used to derive the least-squares problem, thus eliminating the need to compute the least-squares weight vector. This modified structure is advantageous in that it requires only limited input data to accurately describe the external environment. Computer simulation results are presented. >