Abstract
The theoretical advantages of ground-satellite optical communication systems can only be exploited, if means can be found to circumvent the limitations due to atmospheric effects. Atmospheric turbulence dominates the analysis and design of these links. The effects of atmospheric turbulence on the performance of ground-satellite optical communication systems and possible countermeasures are investigated in this thesis. The design and analysis of any optical system operating in the atmosphere requires empirical investigations of atmospheric turbulence conditions at the system's location. Stellar observations provide a useful and convenient means for this purpose in the case of ground-satellite optical communications. The available techniques are reviewed. The experimental setup for a particular technique is described and initial results are presented. We were involved in the first ever ground-satellite optical communication experiments conducted between Japanese Engineering Test Satellite VI and a ground station in Tokyo. One issue which has not yet been satisfactorily resolved is the probability density function (PDF) of intensity fluctuations. It was theoretically shown that the PDF approaches a negative exponential in the very strong turbulence region. Experimental evidence is presented in support of this prediction. The ETS-VI experiment results also confirm that too large a beam size can have significantly deleterious effects on fading performance. Early analyses predicted drastic reductions in uplink on-axis scintillation variance with increasing beam size. As the beam size is increased, the scintillation variance gradient off the beam centre becomes large, and eventually the limitation of the first-order theory is exceeded. An explicit limit on the beam size is identified in this thesis: the beam radius must not exceed a third of the coherence scale. Analyses also predict that appropriately converging the beam results in less scintillation compared to a collimated beam. During the ETS-VI experiment we were not able confirm this prediction. Supplements to the first-order theory also suggest that converging beams behave very similarly to collimated beams. This makes the uplink beam size the single most important adjustable parameter. Possible countermeasures to the atmospheric turbulence effects are identified and reviewed separately for the downlink and the uplink. It is emphasized that uplink transmitter beam size is a crucial design parameter and its optimum value changes continuously according to changing turbulence conditions along the propagation path. A previous study concluded that the optimum beam size is of the order of the coherence scale. It is shown that the optimum size is critically dependent on beam wander and pointing accuracy, and can in fact be much smaller. A novel countermeasure is proposed in which the uplink transmitter beam size is controlled in real time in response to measured turbulence parameters to maximize mean intensity and minimize fluctuations at the satellite receiver.