A neuro-fuzzy computing technique for modeling laser-diode nonlinearity in a radio-over-fibre link

摘要

Neuro-fuzzy (NF) models are well-known robust learning systems that combine the advantages of fuzzy sets and neurocomputation theory. This article presents the application of the NF approach to the computation of 3(rd)-order intermodulation power levels in microwave-intensity-modulated GaAlAs laser diodes for radio-over-fibre communications. The excellent agreement between the computed results and experimental measurements validates the presented NF approach. (c) 2005 Wiley Periodicals, Inc.