• DocumentCode
    761584
  • Title

    Group codes generated by finite reflection groups

  • Author

    Mittelholzer, Thomas ; Lahtonen, Jyrki

  • Author_Institution
    Inst. for Signal & Inf. Process., Swiss Federal Inst. of Technol., Zurich, Switzerland
  • Volume
    42
  • Issue
    2
  • fYear
    1996
  • fDate
    3/1/1996 12:00:00 AM
  • Firstpage
    519
  • Lastpage
    528
  • Abstract
    Slepian-type group codes generated by finite Coxeter groups are considered. The resulting class of group codes is a generalization of the well-known permutation modulation codes of Slepian (1965), it is shown that a restricted initial-point problem for these codes has a canonical solution that can easily be computed. This allows one to enumerate all optimal group codes in this restricted sense and essentially solves the initial-point problem for all finite reflection groups. Formulas for the cardinality and the minimum distance of such codes are given. The new optimal group codes from exceptional reflection groups that are obtained achieve high rates and have excellent distance properties. The decoding regions for maximum-likelihood (ML) decoding are explicitly characterized and an efficient ML-decoding algorithm is presented. This algorithm relies on an extension of Slepian´s decoding of permutation modulation and has similar low complexity,
  • Keywords
    group theory; maximum likelihood decoding; modulation coding; ML decoding algorithm; canonical solution; cardinality; decoding regions; distance properties; finite Coxeter groups; finite reflection groups; group codes; initial point problem; maximum likelihood decoding; minimum distance; optimal group codes; permutation modulation codes; AWGN; Additive white noise; Euclidean distance; Information processing; Mathematics; Maximum likelihood decoding; Modulation coding; Reflection; Signal processing;
  • fLanguage
    English
  • Journal_Title
    Information Theory, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    0018-9448
  • Type

    jour

  • DOI
    10.1109/18.485721
  • Filename
    485721