Energy based transmission optimisation in nonlinear Fourier domain

The systems employing nonlinear Fourier transform (NFT) as a method of nonlinearity mitigation have recently become the subject of intensive study (see e.g. [1-4] and references therein). In particular the so-called nonlinear inverse synthesis (NIS) method was proposed in [2]. Within this method the data is modulated using the continuous part of the NFT spectrum. Then the time domain waveform is generated using the inverse NFT (INFT) before being launched into the fiber. At the receiver NFT operation is performed to retrieve the continuous NFT spectrum (Fig. 1a, see [4]). We consider here the model channel when signal evolution is described by the lossless nonlinear Schrodinger equation (NLSE) with AWGN term arising due to ideal distributed amplification [5]. The signal-noise interaction inside the NFT domain was studied in [3] including the noise statistics.