• DocumentCode
    1889674
  • Title

    A sufficient condition for optimality of digital versus analog relaying in a sensor network

  • Author

    Thejaswi, C. ; Cochran, Douglas ; Zhang, Junshan

  • Author_Institution
    Dept. of Electr. Eng., Arizona State Univ., Tempe, AZ
  • fYear
    2009
  • fDate
    18-20 March 2009
  • Firstpage
    202
  • Lastpage
    206
  • Abstract
    This paper compares the efficacy of digital versus analog relaying in a sensor network when the ultimate goal is detection. A simplified scenario is considered in which a binary decision is to be made at a fusion center based on data it receives from a single sensor. The sensor computes a statistic corresponding to its observations and forwards it over a noisy channel to the fusion center. Two relay schemes: ldquoestimate-and-forwardrdquo (analog relaying) and ldquodetect-and-forwardrdquo (digital relaying) are compared, in terms of the ultimate detection performance they support at the fusion center. With this performance criterion, the relative merit of the two schemes is shown to depend on the observation SNR at the sensor and the SNR of the communication link connecting the sensor and the fusion center. A sufficient condition, in terms of these SNRs, for the superiority of digital relaying over analog relaying is derived for this simple network model.
  • Keywords
    AWGN channels; error statistics; wireless sensor networks; additive Gaussian noise channel; analog relaying; detect-and-forward; digital relaying; error probability; estimate-and-forward; fusion center; wireless sensor networks; Digital relays; Event detection; Monitoring; Sensor fusion; Sensor phenomena and characterization; Sensor systems; Statistics; Sufficient conditions; Surveillance; Wireless sensor networks; Hypothesis-testing; analog/digital relays; sensor networks; signal-to-noise ratio;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Sciences and Systems, 2009. CISS 2009. 43rd Annual Conference on
  • Conference_Location
    Baltimore, MD
  • Print_ISBN
    978-1-4244-2733-8
  • Electronic_ISBN
    978-1-4244-2734-5
  • Type

    conf

  • DOI
    10.1109/CISS.2009.5054717
  • Filename
    5054717