We propose an FFT-based Kramers-Kronig algorithm that can adaptively adjust its computational complexity according to the intensity of turbulence. Compared to the conventional algorithm, it can reduce computational energy consumption by more than 60%.
Li et al. (Tue,) studied this question.