• DocumentCode
    1569251
  • Title

    Performance of sinusoidally-distributed dithering for signed-error constant modulus algorithm

  • Author

    Jusak, Jusak ; Hussain, Zahir M.

  • Author_Institution
    Sch. of Electr. & Comput. Eng., R. Melbourne Inst. of Technol., Vic., Australia
  • Volume
    2
  • fYear
    2005
  • Firstpage
    626
  • Abstract
    The complexity of the ubiquitous constant modulus algorithm (CMA) can be reduced by applying a sign operation to the CMA error function, while a sign operator generally can be seen as a 1-bit (two-level) mid-riser quantization. However, in order to preserve the information lost as a result of quantization processes, dithering is commonly used. In this paper, we evaluate the performance of dithered signed-error CMA utilizing a dithering signal that has sinusoidal probability density function. The algorithm has also been applied for equalization in wireless communication channel, which is distorted by multi-path propagation. Simulation results show that dithering using sinusoidal distribution outperforms the dithering using uniform distribution in terms of convergence speed.
  • Keywords
    blind equalisers; computational complexity; error analysis; multipath channels; quantisation (signal); statistical distributions; ubiquitous computing; waveform generators; CMA error function; computation complexity; mid-riser quantization; multipath propagation; signed-error constant modulus algorithm; sinusoidal probability density function; sinusoidally-distributed dithering; ubiquitous constant modulus algorithm; wireless communication channel equalization; Australia; Computer errors; Convergence; Equalizers; Filters; HDTV; Pervasive computing; Probability density function; Quantization; Wireless communication;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Technology and Applications, 2005. ICITA 2005. Third International Conference on
  • Print_ISBN
    0-7695-2316-1
  • Type

    conf

  • DOI
    10.1109/ICITA.2005.214
  • Filename
    1489037