Regresar

Phase-noise Compensation for QPSK-RoF-OFDM Signals with the Extreme Learning Machine Algorithm for Multilayer Perceptron

Abstract:

Radio-over-fiber orthogonal frequency division multiplexing (RoF-OFDM) technology is negatively affected by laser phase noise and chromatic dispersion optical fiber. These impairments normally generate inter-carrier interference (ICI). An extreme learning machine (ELM)-based receiver for RoF-OFDM schemes is proposed to diminish the ICI effect. The introduced ELM method, composed of various hidden layers, is designed to real-time perform the phase-noise estimation to the received signal, based on the adoption of the pilot subcarriers as the training set, as well as the ELM benefits: good generalization and speed learning. Numerical results show that by appropriately setting the number of hidden nodes, the ELM with three hidden nodes achieves a lower bit error rate (BER) than the benchmarking pilot-assisted equalization and the rest of the ELM approaches reported in the literature.