DocumentCode
2318869
Title
Selection diversity for wireless optical communications with non-coherent detection without CSI
Author
Moradi, Hassan ; Refai, Hazem H. ; LoPresti, Peter G. ; Atiquzzaman, Mohammed
Author_Institution
Electr. & Comput. Eng., Univ. of Oklahoma, Tulsa, OK, USA
fYear
2010
fDate
6-10 Dec. 2010
Firstpage
1010
Lastpage
1014
Abstract
The availability of Channel State Information (CSI) at the receiver side of a Free Space Optical (FSO) link is required for optimal detection and diversity combining. Consequently, providing instantaneous CSI knowledge introduces challenges to receiver design and an increase in estimation error and bandwidth waste. In this work we investigate two selection-combining diversity schemes and their achieved gains when perfect instantaneous CSI is not available. An average SNR estimation value at the receiver is used for branch selection, and signal combining is performed with branch balance and unbalance. Channel fading is modeled as a lognormal distribution with spatially correlated samples. BER performance and outage probabilities are analytically characterized. Further, analytical results are verified using computer simulation.
Keywords
diversity reception; error statistics; fading channels; log normal distribution; optical links; optical receivers; BER performance; FSO link; SNR estimation; bandwidth waste; branch balance; branch selection; channel fading; channel state information; estimation error; free space optical link; lognormal distribution; noncoherent detection; optimal detection; outage probability; receiver; selection-combining diversity; signal combining; wireless optical communication; Channel state information (CSI); Correlated channel; Equal gain combining (EGC); Free Space Optics (FSO); Selection combining (SC); Spatial diversity;
fLanguage
English
Publisher
ieee
Conference_Titel
GLOBECOM Workshops (GC Wkshps), 2010 IEEE
Conference_Location
Miami, FL
Print_ISBN
978-1-4244-8863-6
Type
conf
DOI
10.1109/GLOCOMW.2010.5700086
Filename
5700086
Link To Document