Analysis of Channel Codes using Machine Learning for FSO Communication System

Perfect channel estimate is not a recommended communication method since it is exceedingly difficult, costly, and time/power intensive. The goal of this work is to find a novel, low-cost, deep learning-based solution. A number of novel deep learning and traditional structure combinations are introduced, examined, and contrasted throughout all atmospheric turbulence regimes, from mild to high., The results show that deep learning is resistant to changes in air turbulence and may perform sufficiently near to the ideal channel estimation approach. The suggested deep learning-based solutions have good performance, minimal complexity, and low cost. As a result, they are advised for mobile communication systems’ channel estimate. Since customers utilize small mobile devices as transceivers, these systems should provide them with advantageous and affordable services. These devices must be inexpensive, simple, and low power users. This work aims to replace deep learning-based channel estimation with a traditional, costly, and difficult technique. A more extensive view on channel codes in various machine learning (ML)-based Free Space Optics (FSO) communication systems studied in this research.