Building upon an 8-tap delayed silicon-on-insulator (SOI) photonic neural network (PNN) pre-equalizer validated at 10-GHz bandwidth orthogonal frequency-division multiplexing (OFDM) intensity-modulation direct detection (IMDD), we experimentally probe the operational boundaries of this architecture by scaling to 20-GHz bandwidth operation and 16-QAM modulation. We quantify performance across 5/10/20-GHz OFDM bandwidths and QAM4/QAM16, approaching the sampling limit imposed by the fixed 25-ps tap spacing. In this way, we experimentally demonstrate its applicability to preserve the signal integrity near the Nyquist limit with improvements in all tested configurations.
Versatile neuromorphic photonic processor for next-generation optical communication systems / Silva, L.C.B., Marciano, P.R.N., Staffoli, E., Segatto, M.E.V., Pavesi, L.. - In: OPTICS LETTERS. - ISSN 0146-9592. - 51:15(2026), pp. 4208-4208. [10.1364/ol.605532]
Versatile neuromorphic photonic processor for next-generation optical communication systems
Staffoli, Emiliano;Pavesi, Lorenzo
2026-01-01
Abstract
Building upon an 8-tap delayed silicon-on-insulator (SOI) photonic neural network (PNN) pre-equalizer validated at 10-GHz bandwidth orthogonal frequency-division multiplexing (OFDM) intensity-modulation direct detection (IMDD), we experimentally probe the operational boundaries of this architecture by scaling to 20-GHz bandwidth operation and 16-QAM modulation. We quantify performance across 5/10/20-GHz OFDM bandwidths and QAM4/QAM16, approaching the sampling limit imposed by the fixed 25-ps tap spacing. In this way, we experimentally demonstrate its applicability to preserve the signal integrity near the Nyquist limit with improvements in all tested configurations.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione
