Analysis of polarization drift of optical signals over deployed aerial-inground fiber connections
AI Breakdown
Get a structured breakdown of this paper — what it's about, the core idea, and key takeaways for the field.
Abstract
Polarization measurements of a classical 1550-nm signal are collected and analyzed on 15-km hybrid aerial-inground fiber connections over 11 months. The spectral area and spectral moments9 of mHz-resolution Fast-Fourier-Transform (FFT) of these measurements are extracted, and related to temperature, humidity, wind speed, and time of day. Spectral area correlations show a strong11 diurnal structure: daytime maxima align with temperatures/wind speed peaks and humidity dips, with lower levels during the night. These diurnal patterns also show seasonality, with higher13 mean and variance in summer than winter. A random forest regressor is used to estimate FFT features from environmental measurements, informed by a theoretical model