For optical links through deep turbulence, closed-loop adaptive optics (AO) can facilitate more efficient communication schemes, such as those based on coherent detection and optically pre-amplified receivers. Perturbation-based wavefront correction algorithms, such as stochastic parallel gradient descent (SPGD), are promising candidates for low size, weight, and power consumption (SWaP) alternatives to conventional AO based on direct wavefront sensing. However, limited actuator bandwidth combined with poor convergence rate can constrain the effective AO refresh rate, and degrade the performance when multiple atmospheric modes need correction. Here, we derive and test a new, generalized, non-stochastic, modal wavefront correction algorithm that utilizes either time- or frequency-division to correct multiple modes simultaneously. Using an end-to-end AO simulation, we show the new approach can relax the actuator bandwidth requirement by up to a factor of 8 in comparison to SPGD. Finally, we describe a hardware testbed that is being used to validate the developed approaches.
|