• DocumentCode
    2804654
  • Title

    Sufficient condition for invertibility of square FIR MIMO systems

  • Author

    Wahls, Sander ; Boche, Holger

  • Author_Institution
    Lehrstuhl fur Informationstheorie und Theor. Informationstechnik, Tech. Univ. Berlin, Berlin, Germany
  • fYear
    2010
  • fDate
    14-19 March 2010
  • Firstpage
    3690
  • Lastpage
    3693
  • Abstract
    We derive a sufficient condition for a square FIR MIMO system to have a causal and stable IIR inverse. The condition requires that the spectral norm of the normalized channel impulse response (i.e., the first tap is the identity matrix) is below a certain bound. Intuitively, this means that the system has a strong first tap. This condition often is easier to check than the usual minimum-phaseness, where the roots of the systems determinant have to be computed. Simple approximations of the bound are found. Furthermore, we also give a negative result: the Wiener Filter, which approximates the inverse under low noise conditions, nevertheless always is non-causal. We apply our results to two inversion problems with causality and stability constraint. These problems arise in oversampled noise-shaping subband coding and residual interference cancellation in precoded systems, respectively.
  • Keywords
    FIR filters; IIR filters; MIMO communication; Wiener filters; encoding; interference suppression; transient response; IIR inverse; Wiener filter; causality constraint; channel impulse response; oversampled noise-shaping subband coding; residual interference cancellation; square FIR MIMO systems; stability constraint; Feedback loop; Filter bank; Finite impulse response filter; IIR filters; Interference constraints; MIMO; Noise cancellation; Stability; Sufficient conditions; Wiener filter; Causality; Equalizers; IIR digital filters; MIMO Systems; Stability;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Acoustics Speech and Signal Processing (ICASSP), 2010 IEEE International Conference on
  • Conference_Location
    Dallas, TX
  • ISSN
    1520-6149
  • Print_ISBN
    978-1-4244-4295-9
  • Electronic_ISBN
    1520-6149
  • Type

    conf

  • DOI
    10.1109/ICASSP.2010.5495876
  • Filename
    5495876