Split-step reconstruction technique for the analysis of soliton propagation

Abstract

ABSTRACT A Split-Step Reconstruction Technique (SSRT) is proposed to enhance the computational accuracy of FMAT. The mainadvantage of using SSRT is that the propagation error due to the split-step iteration approach can be eliminated. In addition,an adaptive mesh control algorithm, with which the allocation of mesh size depends on the gradient of soliton, is utilized toimprove the computational efficiency. It is shown that the calculation error can be reduced significantly when comparedwith the FMAT.Keywords: Optical solitons, modeling, optical fiber communication 1. INTRODUCTION Inverse Scattering Method (ISM), Split-Step Fourier Method (SSFM), Fourier Series Analysis Technique (FSAT) andFuzzy Mesh Analysis Technique (FMAT) are methods available to study the propagation phenomena of solitons in anoptical fiber [1 ]—[7]. The ISM gives an analytical solution to the propagation of soliton in a lossless fiber. For higher ordersolitons, the complexity and difficulty offinding exact solutions using ISM will be greatly increased. SSFM requires largenumber of sampling points and depends heavily on the calculation of fast Fourier transform. In order to minimize thiscumulative errors, FSAT is developed to operate entirely in frequency domain. This technique is very efficient when therequired number of sampling points is small. However, the computational efficiency of FSAT reduces with an increase ofsampling points. Recently, FMAT is proposed for the enhancement ofcomputational speed and efficiency. This technique iscapable of analyzing the propagation phenomena of high power soliton, pulse compression and soliton interaction in anefficient manner.In this paper, a Split-Step Reconstruction Technique (SSRT) is proposed to further enhance the computational accuracy ofFMAT. The main advantage of using SSRT is that the propagation error due to the split-step iteration approach can beeliminated. In addition, an adaptive mesh control algorithm, with which the allocation ofmesh size depends on the gradientof soliton, is utilized to improve the computational efficiency.