Abstract

This work investigates a probabilistic model for splice loss in single mode optical fibers. We derive the probability density function for loss values as a function of lateral and angular misalignment. We then use observed data to estimate these model parameters; both Bayesian and maximum likelihood estimation procedures are described. These estimates can then be used to provide an indication of the relative importance of various loss mechanisms. Alternatively, if one is given values for maximum lateral and angular misalignment, our results allow for predictions of expected distribution of loss values. An overall goal of this paper is to demonstrate that, beyond the mean and variance of splice loss, there is significant information in the shape of the distribution of values. A second goal is to understand the trade-off between the number of splice loss measurements and the confidence in estimates of parameters in the splice loss model.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call