%675$&7When designing free-space optics systems, one key issue is to assess the impact of scintillations and to find an appropriate link margin to cope with atmospheric fading. Huge effort is spent to find mathematical models to describe laser beam propagation through the atmosphere. However, these models are quite cumbersome to use for the communications engineer. On the other hand, there are empirical models that try to describe the influence of scintillations and other system parameters in a simple and easy to use manner. Nevertheless, they are empirical and not based on theory. This paper is intended to close the gap between mathematical theory and empirical models.Therefore, a simple yet accurate receiver model is introduced. Based on turbulence theory and using the recently proposed convolution method assuming independent sub-aperture intensities, probability distributions of the received power are derived. Aperture averaging as well as multiple transmitter systems can be described this way. Power penalties are found by numerically calculating the resulting bit error probabilities for varying mean values of received power. Finally a model for appropriate link margins under different atmospheric conditions, taking transmitter diversity and aperture averaging into account, is derived and compared to empirical models..H\ZRUGV FSO, scintillations, scintillation loss, receiver model
,1752'8&7,21Optical free-space communications systems greatly suffer from intensity fluctuations (scintillations) at the receive aperture due to atmospheric turbulence. Varying optical input power to the receiver leads to varying bit error probabilities. Deep fades with respectively high bit error probabilities may significantly deteriorate the over-all system performance. Scintillation effects are mitigated by use of multiple transmit and/or receive apertures and by relying on aperture averaging over large receive apertures. To determine the system performance two basic things have to be known: the probability density function (PDF) of received power and a receiver model giving a relation between received power and bit error probability. Whereas the latter one depends on the characteristics of the receiver circuitry, the former is influenced by a number of system and atmospheric parameters: The link distance L, the accumulated number N of transmit and receive apertures, the diameter of the receive aperture(s), the according aperture averaging factor and finally the turbulence strength conveniently expressed through the parameter C n 2 . The necessary PDF of received power can be derived from arbitrary models of intensity PDF by using the convolution method as described in [1]. For the weak fluctuation regime a lognormal distribution can be assumed for the intensity fluctuations.In the following we derive a receiver model valid for IM/DD schemes using PIN or APD photo diodes. Subsequently we review the weak fluctuation theory and explain the convolution method for deriving receive power distributions. Finally we assess the bit error ...