학술논문

Reliable Free-Space Optical Communication System Performance for Matching Multi-level Customer Needs: Using Hybrid Modulation System with Deep Reinforcement Learning
Document Type
Conference
Source
2022 IEEE Symposium on Computers and Communications (ISCC) Computers and Communications (ISCC), 2022 IEEE Symposium on. :1-7 Jun, 2022
Subject
Communication, Networking and Broadcast Technologies
Computing and Processing
Multiplexing
Free-space optical communication
Communication systems
Atmospheric modeling
System performance
Computational modeling
Reinforcement learning
Free-space optical communication (FSO)
at-mospheric turbulence
deep learning (DL)
wavelength division multiplexing (WDM)
pointing error
multiple-input-single-output (MISO)
single-input-multiple-output (SIMO)
multiple-input-multiple-output (MIMO)
bit-error-rate (BER)
signal-to-noise ratio (SNR)
equipment mean life-span
Language
ISSN
2642-7389
Abstract
In the recent decades, free-space optical (FSO) communication technology has gained significant importance owing to its promising unique features: high user capacity, license-free spectrum, ease and quick deploy-ability. However, the performance of FSO communication systems depends on the uncontrollable terrestrial atmospheric effects. The second main challenge is that the FSO system performance degrades as a line-of-Sight (LoS) technology due to the misalignment transmitter and receiver. The third challenge is the consideration of time value of money, which is central to most engineering economic analyses in employing communication systems. The opportunity cost of making one choice over another must also be considered. This paper presents a proposed FSO system design model for mitigating the three performance challenges we mentioned. The results show an enhancement by 83.34% in system efficiency in case of moderate scintillation (using gamma-gamma model). This proposed hybrid model is proved to be applicable to any atmospheric channel conditions.