• DocumentCode
    1732577
  • Title

    Soft-output sphere detection for coded Unique Word OFDM

  • Author

    Onic, Alexander ; Schenk, Andreas ; Huemer, Mario ; Huber, Johannes B.

  • Author_Institution
    Inst. of Networked & Embedded Syst., Alpen-Adria-Univ. Klagenfurt, Klagenfurt, Austria
  • fYear
    2012
  • Firstpage
    138
  • Lastpage
    142
  • Abstract
    Unique Word (UW) OFDM is an attractive alternative to conventional cyclic-prefix OFDM as it offers the potential to apply more sophisticated detection schemes, resulting in improved performance both in uncoded as well as coded transmission. In this paper, we study soft-output detection schemes suited for coded UW-OFDM transmission. In particular, we present an efficient implementation of near-optimum soft-output detection based on the soft-output sphere decoder (SOSD) algorithm. The performance and complexity of this approach is compared to a low-complexity linear detector (linear minimum mean squared error, LMMSE), which uses exact statistics of the residual noise for computation of the soft output. The conducted numerical simulations emphasize that the SOSD constitutes the decoding reference for uncoded as well as coded UW-OFDM transmission. Nevertheless, the LMMSE offers a good performance trade-off at a constant complexity, while the SD´s complexity is variable.
  • Keywords
    MIMO communication; OFDM modulation; decoding; least mean squares methods; LMMSE; SOSD algorithm; UW-OFDM transmission; coded transmission; decoding reference; linear minimum mean squared error; low-complexity linear detector; residual noise; soft-output sphere decoder; soft-output sphere detection; uncoded transmission; unique word;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signals, Systems and Computers (ASILOMAR), 2012 Conference Record of the Forty Sixth Asilomar Conference on
  • Conference_Location
    Pacific Grove, CA
  • ISSN
    1058-6393
  • Print_ISBN
    978-1-4673-5050-1
  • Type

    conf

  • DOI
    10.1109/ACSSC.2012.6488976
  • Filename
    6488976