![Exploring methods to forecast the intensity scintillation in free space optical communication using a deep learning approach](https://ecommons.udayton.edu/stander_posters/3909/thumbnail.jpg)
Exploring methods to forecast the intensity scintillation in free space optical communication using a deep learning approach
Presenter(s)
Mohammad Albaqer Hammid Jwaid Al Ghezi
Files
Description
Free space optical communication (FSO) is an essential technology that uses optical bandwidthto transmit data through the air, it can transmit up to 2.5Gbps through a secure channel.However, there are several challenges an FSO channel encounters, one of which isatmospheric turbulence. Atmospheric turbulence can degrade the optical signal due to effects,such as intensity scintillation and beam wandering. The scintillation index is an often-usedmetric measuring the normalized intensity variance. It can be measured using a scintillometer.However, it is not possible to measure the scintillation index in all locations and at all times.In this work, a machine learning algorithm has been optimized to forecast the scintillation index.Meteorological data, such as air temperature, humidity, and wind speed, is obtained togetherwith the scintillation index at an experiment along a 7 km propagation path in Dayton, OH. Thedata is divided into four equal parts corresponding to the four seasons and the data in eachseason is divided into training and validation data. Long-short-term memory (LSTM) modelshave been optimized and tuned to forecast the scintillation index. The mean absolute error(MAE) is used to compare the predicted scintillation index with the measured scintillation indexand the adaptive moment estimation (Adam) optimizer is used to update the trainableparameters to fit the scintillation. The training process is performed with different LSTM modelson the training data for each season and the performance of the model is measured using thevalidation data for the corresponding season. The LSTM model predicts the scintillation indexwith weighted average MAE around 0.07 for all seasons.
Publication Date
4-19-2023
Project Designation
Graduate Research
Primary Advisor
Miranda van Iersel
Primary Advisor's Department
Electro-Optics and Photonics
Keywords
Stander Symposium, School of Engineering
Institutional Learning Goals
Scholarship
Recommended Citation
"Exploring methods to forecast the intensity scintillation in free space optical communication using a deep learning approach" (2023). Stander Symposium Projects. 2908.
https://ecommons.udayton.edu/stander_posters/2908
![Exploring methods to forecast the intensity scintillation in free space optical communication using a deep learning approach](https://ecommons.udayton.edu/stander_posters/3909/thumbnail.jpg)
Comments
Presentation: 2:40-3:00 p.m., Jessie Hathcock Hall 180