Multi-layer perceptron as equalisers for multilevel pulse amplitude modulation scheme in SI-POF

To improve performance in step-index plastic optical fiber (SI-POF), spectral efficient modulation Scheme like multilevel pulse amplitude modulation (M-PAM) with equalisers have recently been studied. The adaptive equalisers used for POF in recent literature are the conventional linear transversal equalisers (LTE) and decision feedback equaliser (DFE). In this paper, we investigate the performance of multi-layer perceptron (MLP), which is a feed-forward neural network architecture, as an adaptive equaliser for minimizing intersymbol interference (ISI) in SI-POF using M-PAM schemes. The bit-error rate (BER) performance for the MLP equaliser is compared with that of conventional equaliser (LTE and DFE) for theoretical POF channels with Gaussian Profile. In comparison to LTE and DFE, we show in this paper that under severe ISI, MLP achieves better BER performance.