Enhancing Complex Light Beam Propagation in Turbulent Atmosphere with Active Convolved Illumination
Active Convolved Illumination is a way to correct for the distortions coming from a channel in complex media, by adding an auxilliary source that counteracts the channel distortions while being not present at the output.
Using band limited modes in k-space with orbital angular momentum at the source,they are sent through the channel and using the measurments on the receiving side the modes are reconstructed in a distorted way. By comparing the phase differenes between true and reconstructed modes a "Correlation Injected Source" (CIS) is created that represents a signal that would correct for the channel in the bandwidth of the corresponding mode. This is then covolved with the corresponding OAM mode to create a correlation injected oam mode.
The modified source signal is the superposition of the convolutions of the original signal with the correlation injected oam modes that cover the relevant signal bandwidth.
This new signal that has been convolved with the CIS beams that have corrected for the distortions in each of their specific bandwidths can be sent through and the reconstruction of this singal can be deconvolved with specific operators to restore the power spectral density of the signal.
The convolution with CIS to form a new signal can be representd as adding a an auxilliary signal to acheive the same effect.
My understanding is that the main point is to the fact that you can correct for the channel conditions in each band, using the OAM modes, so that you can then take that correction an apply it to an unknown source spatial distribution using convolution.
What they show in the results is adding the auxilliary signals to the source and recovering just the original source signal with minimal distortion.
- need to read more of the results
Questions
How are you supposed to create the OAM modes? Can the processing be done fast enough with the OAM modes before the channel changes? I believe this paper assumes that the channel is taken as unchanging for the time between measurng the errors in the modes and then sending the ACI signal.
If the channel is strongly non-linear, does the band limited modes make sense? There would be a more complicated relationship and spectral leakage.