GMI-maximizing constellation design with Grassmann projection for parametric shaping

Toshiaki Koike-Akino, David S. Millar, Kieran Parsons, Keisuke KojimaOFC 2016 · OpenReview

Abstract

We introduce a new geometric shaping to optimize signal constellations maximizing generalized mutual information for both binary and nonbinary coding. With parametric projection of hyper cubes, we can perform seamless constellation shaping with near 1 dB gain.
BibTeX
@inproceedings{koikeakino2016gmi,
  author = {Toshiaki Koike-Akino and David S. Millar and Kieran Parsons and Keisuke Kojima},
  title = {{GMI-maximizing constellation design with Grassmann projection for parametric shaping}},
  booktitle = {OFC},
  year = {2016},
  url = {https://openreview.net/forum?id=RGOkmLu5LJC},
}

Loading related content...

Figure