• DocumentCode
    890643
  • Title

    On compound channels with side information at the transmitter

  • Author

    Mitran, Patrick ; Devroye, Natasha ; Tarokh, Vahid

  • Author_Institution
    Div. of Eng. & Appl. Sci., Harvard Univ., Cambridge, MA, USA
  • Volume
    52
  • Issue
    4
  • fYear
    2006
  • fDate
    4/1/2006 12:00:00 AM
  • Firstpage
    1745
  • Lastpage
    1755
  • Abstract
    Costa has proved that for noncausally known Gaussian interference at a power constrained transmitter communicating over an additive white Gaussian noise channel there is no capacity loss when compared to a scenario where interference is not present. For the case of a transmitter communicating over a quasistatic (i.e., nonergodic) fading channel, his method does not apply. In this correspondence, we derive upper and lower bounds on the capacity of compound channels with side information at the transmitter, first for finite alphabet channels and then, based on this result, for channels on standard alphabets (this includes real alphabets). For the special case of a degenerate compound channel with only one possible realization, our bounds are equivalent to the well-known capacity with side-information formula of Gel´fand and Pinsker. For the quasistatic fading channel, when fading is Ricean, we suggest a scheme based on our lower bound for which the performance is found to be relatively good even for moderate K-factor. As K→∞, the uncertainty on the channel vanishes and our scheme obtains the performance of dirty paper coding, namely that the interference is perfectly mitigated. As K→0, the proposed scheme treats the interferer as additional noise. These results may be of importance for the emerging field of cognitive radios where one user may be aware of another user´s intended message to a common receiver, but is unaware of the channel path gains.
  • Keywords
    AWGN channels; Rician channels; channel capacity; radio transmitters; radiofrequency interference; software radio; Gaussian interference; additive white Gaussian noise channel; cognitive radio; compound channel; dirty paper coding; finite alphabet channel; moderate K-factor; quasistatic Ricean fading channel; side information; transmitter communication; Additive noise; Additive white noise; Channel capacity; Cognitive radio; Fading; Gaussian noise; Interference constraints; Propagation losses; Radio transmitters; Uncertainty; Cognitive radios; Gel´fand–Pinsker channel; compound channels; dirty paper coding; fading channels; side information;
  • fLanguage
    English
  • Journal_Title
    Information Theory, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    0018-9448
  • Type

    jour

  • DOI
    10.1109/TIT.2006.871044
  • Filename
    1614102