• DocumentCode
    2264752
  • Title

    Perfect reconstruction DFT modulated oversampled filter bank transceivers

  • Author

    Rahimi, Siavash ; Champagne, Benoit

  • Author_Institution
    Dept. of Electr. & Comput. Eng., McGill Univ., Montreal, QC, Canada
  • fYear
    2011
  • fDate
    Aug. 29 2011-Sept. 2 2011
  • Firstpage
    1588
  • Lastpage
    1592
  • Abstract
    This paper proposes a novel method for the design of perfect reconstruction (PR), discrete Fourier transform (DFT) modulated oversampled filter banks (FB) for application in multi-carrier transceiver systems. The PR property is guaranteed by enforcing a paraunitary constraint on the polyphase matrix of the transmit or receive sub-systems. The desired polyphase matrix is obtained via embedding of lower dimensional paraunitary building blocks, each expressed in terms of a limited set of design parameters through a factorization based on Givens rotations. These parameters are adjusted to minimize the stop-band energy of the subband filters and thus improve their spectral containment. The performance of the proposed FB is investigated in a multi-carrier transceiver application, where it is compared with OFDM and other recently proposed FB structures. Numerical results show that the proposed approach can lead to significant reduction in the bit error rate (BER), as compared to the benchmark approaches, when used in the presence of narrowband interference or frequency offset.
  • Keywords
    channel bank filters; discrete Fourier transforms; error statistics; matrix algebra; transceivers; FB; FB structures; Givens rotations; PR property; bit error rate; discrete Fourier transform; multicarrier transceiver systems; paraunitary constraint; perfect reconstruction DFT modulated oversampled filter bank transceivers; polyphase matrix; spectral containment; stop-band energy; Attenuation; Bit error rate; Discrete Fourier transforms; Filter banks; OFDM; Prototypes; Transceivers;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal Processing Conference, 2011 19th European
  • Conference_Location
    Barcelona
  • ISSN
    2076-1465
  • Type

    conf

  • Filename
    7073898