Experimental analysis of the time dynamics of coherent communication through turbulence: Markovianity and channel prediction

Clear air atmospheric turbulence causes significant fading for terrestrial-terrestrial and terrestrial-satellite free space optical communication systems. Typically extra link margin is used to assure link availability and reliability, however this extra margin is an inefficient and expensive use of resources. In this paper, we analyze data collected by an experimental system with a single laser transmitter located 250 meters from two coherent receivers. We first use the data to validate the use of a two-state continuous time Markov process to model outage statistics of the diversity system. In the two-state channel model, symbols received during an outage are assumed to be lost, and symbols received during a non-outage are assumed to be received correctly. This channel model can be used to analyze the performance of the transport layer. Next, we use statistical and spectral analysis techniques to create a linear prediction model for signal attenuation for both the single-receiver and diversity systems. The prediction model is an optimal estimator that predicts signal attenuation 1 ms into the future to 1.5 dB accuracy for the single-receiver cases and to 1 dB accuracy for the diversity case. The maximum amount of time the estimator can predict into the future with some confidence is about 5-10 ms. This channel prediction and adaptation can be used to greatly improve the efficiency of free-space optical communication systems in the atmosphere.