Метод демультиплексирования геликоидального волнового фронта

Machine learning methods were used to construct a demultiplexer for helical wave front separation into orthogonal modes. The accuracy of wave front demultiplexing into eight modes at a signal-to-noise ratio of –3 dB is about 95% in a broad range of signal carrier frequencies. For nonstationary parameters of signals, the proposed demultiplexer accuracy exceeds that of the classical correlation method.