BER analysis of optical wireless signals through lognormal fading channels with perfect CSI

Due to inconsistent atmospheric conditions, scattering and scintillation of free space optical (FSO) signal can occur, thus negatively influencing the received signal intensity. The channel is usually modeled as a normalized fading coefficient with additive Gaussian noise. Optimal detection of the received signal is designed based on a decision rule, e.g., Maximum Likelihood (ML), assuming the receiver knows the noise statistics and fading correlation of the channel. This paper briefly deals with analysis on bit error rate (BER) of a wireless optical signal passing through a lognormally distributed fading channel, when perfect knowledge of channel state information (CSI) at the receiver side is available. Two approaches will be presented to provide closed-form expressions for BER. One uses Gauss-Hermite quadrature approximation and the other one is based on power series. While numerical analysis shows a very small approximation error when the Gauss-Hermite approach is considered, the power series approach does not uses any approximation.