Research Article Open Access

On the Pulsewidth Analysis in the Presence of PMD and PDL in Optical Fibers Using Neural Network Algorithm

A. Sivasubramanian and V. C. Ravichandran

Abstract

In long haul networks, the random birefringence induced in the optical fiber leads to a considerable Polarization Mode Dispersion (PMD). Polarization Dependent Loss (PDL) mainly occurs in optical components and depends on the state of polarization of optical signals. The presence of PMD and PDL causes pulsewidth narrowing and the pulsewidth reduction depends on states of polarization at which the input light launched and also the input pulsewidth. A system comprising of a PDL element sandwiched between two PMD elements was considered. This system was characterized using neural network approach. Back propagation algorithm was applied to train the network with four input vectors namely PMD, PDL, input pulsewidth and the angle describing the input states of polarization and one output vector indicating effective squared pulsewidth difference. On analysis, it was found that the pulsewidth reduction was higher for a PMD of 30ps, a PDL of 3.5 and input pulsewidth of 100ps at various (Linear and Circular) input states of polarization with the angle describing the input state of polarization to be |π/4|. Similarly, for a given value of PMD, PDL, input pulsewidth and a specific pulsewidth reduction, the input state of polarization at which the light was to be launched can also be determined using neural network approach.

Journal of Computer Science
Volume 3 No. 4, 2007, 238-241

DOI: https://doi.org/10.3844/jcssp.2007.238.241

Submitted On: 29 November 2006 Published On: 30 April 2007

How to Cite: Sivasubramanian, A. & Ravichandran, V. C. (2007). On the Pulsewidth Analysis in the Presence of PMD and PDL in Optical Fibers Using Neural Network Algorithm . Journal of Computer Science, 3(4), 238-241. https://doi.org/10.3844/jcssp.2007.238.241

  • 2,752 Views
  • 2,010 Downloads
  • 3 Citations

Download

Keywords

  • Polarization mode dispersion
  • polarization dependent loss
  • pulse narrowing
  • neural networks