Integrated Silicon Photonic Reservoir Computing With PSO Training Algorithm for Fiber Communication Channel Equalization

Abstract

An optical channel distortion equalization method based on silicon photonic reservoir computing (PhRC) structure with particle swarm optimization (PSO) algorithm is proposed. The weights training scheme of the photonic readout layer of PhRC is trained by PSO algorithm towards a lower bit error rate (BER) and a better eye diagram, simultaneously. The self-organizing evolutionary PSO algorithm enables fast convergence and iterative optimization for training optical weights. Without the necessary the reservoir signal states obtained from the monitors, it eliminates the noise introduced by the optical monitors. We implement a system-level simulation, including RC structure, training algorithm, and fiber communication construction. Because of the excellent performance of the PSO training algorithm, a BER of 9.15 × 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−5</sup> is obtained with the 25 Gb/s on-off keying (OOK) input signal in the case of 25 km single-mode transmission. This BER result is three orders of magnitude lower than that before the equalization. This equalizer mitigates the processing bandwidth limit and provides a distorted signals equalization method in the optical domain with simple and effective training strategy. This method shows much potential in high speed optical communication channel equalization in the future.