Measurements of Optical Turbulence and Analysis Using Machine Learning
Abstract
Optical turbulence impacts the performance of laser weapons and laser communication by disrupting the focus of the laser beam. It is important to characterize the turbulence along the beam path in order to predict the performance of these systems. Unfortunately, the equipment needed to measure optical turbulence is delicate. A previous thesis found that the turbulence can be estimated using machine learning regression analysis trained on simple atmospheric measurements that can be made with more robust instruments. Machine learning regression analysis is a powerful tool to model complex phenomena with no clear analytical relationship, although extensive data sets are required to train the machine learning model. For this thesis, we measured optical turbulence and various atmospheric parameters (air temperature, humidity, solar flux, etc.) over many months. Using measured atmospheric parameters as inputs, we developed an ensemble of bagged trees regression model with optical turbulence as the response. Overall, this model showed good agreement with the measured values of turbulence. This indicates turbulence could be predicted using these more robust instruments coupled with a machine learning regression model.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2021
- Accession Number
- AD1165008
Entities
People
- Antonios Sklavounos
Organizations
- Naval Postgraduate School