Illiad: a Resource-Efficient Audio Sensing Service Based on the Guided Search Model
Long Le, David Jun, Douglas L. Jones

Citation
Long Le, David Jun, Douglas L. Jones. "Illiad: a Resource-Efficient Audio Sensing Service Based on the Guided Search Model". Talk or presentation, November, 2014; Poster presented at the 2014 TerraSwarm Annual Meeting.

Abstract
To handle various applications' queries, sensing services in the Internet of Things (IoT) must often process a large volume of unstructured data, i.e. audio/video data, generated across both time and space by autonomous sensors. As the IoT expands at an exponential rate, the amount of unstructured data that need processing has become unprecedentedly large, rendering the naive solution that streams all these data to the cloud either infeasible or cost-prohibitive. Enabling sensing services using finite sensing, computing, and storage resources requires a new paradigm. Inspired by the Visual Guided Search model of the human visual detection and search system, we introduce Illiad, a resource-efficient audio sensing service that is based on the Guided Search paradigm. The paradigm is structured around two levels, with a key idea that information in the first, application-agnostic level can be used to guide the deployment of resources for application-specific processing in the second level. This way resources are deployed exactly where they are most useful, and the system remains resource-efficient across all levels. The Illiad service was used to create several web apps that, to the authors' best knowledge, have not been done before. Evaluation shows that, on average, the Illiad audio sensing service achieves a saving ratio of 12 X in term of bandwidth and storage, compared to the naive approach of sending all data to the cloud. In addition, its Android sensing platform only increases the energy consumption by 10%-27% compared to continuous recording.

Electronic downloads

Internal. This publication has been marked by the author for TerraSwarm-only distribution, so electronic downloads are not available without logging in.

Citation formats

• HTML

  Long Le, David Jun, Douglas L. Jones. <a
  href="http://www.terraswarm.org/pubs/429.html"><i>Illiad:
  a Resource-Efficient Audio Sensing Service Based on the
  Guided Search Model</i></a>, Talk or
  presentation,  November, 2014; Poster presented at the <a
  href="http://www.terraswarm.org/conferences/14/annual"
  >2014 TerraSwarm Annual Meeting</a>.

• Plain text

  Long Le, David Jun, Douglas L. Jones. "Illiad: a
  Resource-Efficient Audio Sensing Service Based on the Guided
  Search Model". Talk or presentation,  November, 2014;
  Poster presented at the <a
  href="http://www.terraswarm.org/conferences/14/annual"
  >2014 TerraSwarm Annual Meeting</a>.

• BibTeX

  @presentation{LeJunJones14_IlliadResourceEfficientAudioSensingServiceBasedOnGuided,
      author = {Long Le and David Jun and Douglas L. Jones},
      title = {Illiad: a Resource-Efficient Audio Sensing Service
                Based on the Guided Search Model},
      month = {November},
      year = {2014},
      note = {Poster presented at the <a
                href="http://www.terraswarm.org/conferences/14/annual"
                >2014 TerraSwarm Annual Meeting</a>.},
      abstract = {To handle various applications' queries, sensing
                services in the Internet of Things (IoT) must
                often process a large volume of unstructured data,
                i.e. audio/video data, generated across both time
                and space by autonomous sensors. As the IoT
                expands at an exponential rate, the amount of
                unstructured data that need processing has become
                unprecedentedly large, rendering the naive
                solution that streams all these data to the cloud
                either infeasible or cost-prohibitive. Enabling
                sensing services using finite sensing, computing,
                and storage resources requires a new paradigm.
                Inspired by the Visual Guided Search model of the
                human visual detection and search system, we
                introduce Illiad, a resource-efficient audio
                sensing service that is based on the Guided Search
                paradigm. The paradigm is structured around two
                levels, with a key idea that information in the
                first, application-agnostic level can be used to
                guide the deployment of resources for
                application-specific processing in the second
                level. This way resources are deployed exactly
                where they are most useful, and the system remains
                resource-efficient across all levels. The Illiad
                service was used to create several web apps that,
                to the authors' best knowledge, have not been done
                before. Evaluation shows that, on average, the
                Illiad audio sensing service achieves a saving
                ratio of 12 X in term of bandwidth and storage,
                compared to the naive approach of sending all data
                to the cloud. In addition, its Android sensing
                platform only increases the energy consumption by
                10%-27% compared to continuous recording. },
      URL = {http://terraswarm.org/pubs/429.html}
  }
  

Posted by Long Le on 4 Nov 2014.