• DocumentCode
    297087
  • Title

    A unified table based trellis subset decoder for high speed voice-band modems

  • Author

    Nobakht, Ramin A.

  • Author_Institution
    IBM Corp., Research Triangle Park, NC, USA
  • Volume
    2
  • fYear
    1994
  • fDate
    28 Nov- 2 Dec 1994
  • Firstpage
    677
  • Abstract
    An unified table based trellis subset decoder for the Viterbi algorithm is designed which is capable of decoding any constellation size for any data rate scheme of the CCITT V.32, V32bis, and IBM “V.32ter” full duplex high speed voice-band modem standards with constant number of computations. In addition, no error is introduced as the result of this subset decoding. The data rates include the trellis coded 7200, 9600, 12000, 14400, 16800, and 19200 bps. The number of computations necessary to perform such a task is the same for any given data rate. This results in reductions in the computational complexity up to a factor of 32 times over the existing direct methods. The memory consumption of this algorithm is small and increases proportionally with the data rate
  • Keywords
    Viterbi decoding; computational complexity; modems; telecommunication standards; trellis codes; voice communication; 12000 bit/s; 14400 bit/s; 16800 bit/s; 19200 bit/s; 7200 bit/s; 9600 bit/s; CCITT V.32; CCITT V32bis; IBM V.32ter; Viterbi algorithm; computational complexity reduction; constellation size; data rate; full duplex high speed modem standards; high speed voice-band modems; memory consumption; subset decoding; trellis subset decoder; unified table; Decoding; Delay; Digital signal processing; Euclidean distance; Lips; Markov processes; Modems; Recursive estimation; State estimation; Viterbi algorithm;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Global Telecommunications Conference, 1994. GLOBECOM '94. Communications: The Global Bridge., IEEE
  • Conference_Location
    San Francisco, CA
  • Print_ISBN
    0-7803-1820-X
  • Type

    conf

  • DOI
    10.1109/GLOCOM.1994.512682
  • Filename
    512682