Low-frequency sampling adaptive thresholding for free-space optical communication receivers with multiplicative noise

Abstract

An adaptive thresholding method is presented for optimum detection for optical receivers with large multiplicative noise. The technique uses low frequency sampling of the detected current that enables calculation of the bit means and variances and estimation of the optimum detection threshold. The regime in which this holds is when the sampling frequency is lower than the bit rate but higher than atmospheric turbulence frequency content. Simulations are done with data obtained from the NRL Chesapeake Bay Lasercomm Testbed. The results of simulations comparing BER performance versus sample rate and parameter estimation error will be presented. If the system parameters are characterized in advance with reasonable accuracy, the BER obtained will typically be an order of magnitude improvement over the equal variance threshold (depending on the signal to noise ratio).