Chromatic Distortion Precompensation in OFDM-Based Optical Systems Through an Integrated Silicon Photonic Neural Network

In this paper, we report an experimental demonstration of using an integrated silicon photonic neural network to overcome the chromatic distortion of direct-detection optical systems based on orthogonal frequency-division multiplexing (OFDM). The precompensation is conducted in the optical domain of a 125 km link of standard single-mode fiber (SSMF), i.e., the proposed approach enables OFDM pre-equalization directly in the optical domain instead of electronic distortion compensation that characterizes common OFDM-based systems. Thanks to the intrinsic features of the OFDM multiplexing format, and to the elimination of cyclic prefix provided by our proposal, improvements in spectral and power efficiencies can be achieved. The performance enhancements registered in the conducted experiments show that photonic integrated circuits can be embedded in future OFDM-based commercial transceivers, paving the way for new and promising architectures of intensity modulation and direct-detection optical communication systems.