End-to-End Optimization of Constellation Shaping for Wiener Phase Noise Channels With a Differentiable Blind Phase Search

As the demand for higher data throughput in coherent optical communication systems increases, we need to find ways increase capacity existing and future links. To address spectral efficiencies, apply end-to-end optimization joint geometric probabilistic constellation shaping presence of Wiener phase noise carrier estimation. Our approach follows state-of-the-art bitwise auto-encoders, which require a differentiable implementation all operations between transmitter receiver, including DSP algorithms. In this work, show how modify ubiquitous blind search (BPS) algorithm, popular estimation make it include shaping. By leveraging shaping, are able obtain robust pilot-free modulation scheme improving performance 64-ary by at least 0.1bit/symbol compared square QAM constellations with neural demappers 0.05 bit/symbol previously presented approaches applying only