Journal
JOURNAL OF LIGHTWAVE TECHNOLOGY
Volume 41, Issue 12, Pages 3797-3815Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/JLT.2023.3272011
Keywords
Artificial intelligence; coherent detection; computational complexity; FPGA; neural network hardware; nonlinear equalizer; recurrent neural networks
Ask authors/readers for more resources
In this work, FPGA realization of neural network-based equalizers for compensating nonlinearity in coherent optical transmission systems is demonstrated. The models are converted from Python libraries to FPGA synthesis and implementation. Different hardware implementation alternatives for nonlinear activation functions are reviewed. Performance comparison, implementation analysis, and hardware complexity are discussed. The experiments show that the biLSTM+CNN equalizer achieves similar results to DBP and provides a Q-factor gain compared to the baseline. Approximations of activation functions using Taylor series, piecewise linear, and look-up table are evaluated in terms of Q-factor and hardware utilization.
In this work, we demonstrate the offline FPGA realization of both recurrent and feedforward neural network (NN)-based equalizers for nonlinearity compensation in coherent optical transmission systems. First, we present a realization pipeline showing the conversion of the models from Python libraries to the FPGA chip synthesis and implementation. Then, we review the main alternatives for the hardware implementation of nonlinear activation functions. The main results are divided into three parts: a performance comparison, an analysis of how activation functions are implemented, and a report on the complexity of the hardware. The performance in Q-factor is presented for the cases of bidirectional long-short-term memory coupled with convolutional NN (biLSTM + CNN) equalizer, CNN equalizer, and standard 1-StpS digital back-propagation (DBP) for the simulation and experiment propagation of a single channel dual-polarization (SC-DP) 16QAM at 34 GBd along 17 x 70 km of LEAF. The biLSTM+CNN equalizer provides a similar result to DBP and a 1.7 dB Q-factor gain compared with the chromatic dispersion compensation baseline in the experimental dataset. After that, we assess the Q-factor and the impact of hardware utilization when approximating the activation functions of NN using Taylor series, piecewise linear, and look-up table (LUT) approximations. We also show how to mitigate the approximation errors with extra training and provide some insights into possible gradient problems in the LUT approximation. Finally, to evaluate the complexity of hardware implementation to achieve 200G and 400G throughput, fixed-point NN-based equalizers with approximated activation functions are developed and implemented in an FPGA.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available