Selecting higher wireless connection capacity routes with/without predicting congestion in access networks

Tutomu Murase, Gen Motoyoshi, Kazuhide Sonoda, Jiro Katto

Research output: Contribution to journalArticlepeer-review


It is important for multimedia mobile users to obtain sufficient resources and a stable capacity of resources. Network resources such as available bandwidth for mobile users, however, fluctuate depending on the level of background traffic. The deployment of several access points enables mobile users to access more network resources on the way to their destination by taking the longcut route rather than the shortcut route. These users pay an extra cost to take the longcut route but obtain greater gains in throughput over the shortcut route. Previous research has shown that the gain-to-cost ratio can be more than double for the longcut route. No considerations, however, have been made for the case in which resource fluctuation is due to network congestion. This paper investigates the quality of service (QoS) for a longcut route with fewer resources (real resources) compared with that for one with the total resources (ideal resources) in the longcut route calculation. The results show that on average, users have gains of approximately double in real resources than their gains in ideal resources. This finding indicates that users can greatly improve their gains if they can accurately predict the resource fluctuation.

Original languageEnglish
Pages (from-to)345-352
Number of pages8
JournalITE Transactions on Media Technology and Applications
Issue number4
Publication statusPublished - 2014


  • Cognitive radio
  • Congestion
  • Mobile user
  • Resource allocation
  • Route planning
  • Wireless LANs

ASJC Scopus subject areas

  • Signal Processing
  • Media Technology
  • Computer Graphics and Computer-Aided Design


Dive into the research topics of 'Selecting higher wireless connection capacity routes with/without predicting congestion in access networks'. Together they form a unique fingerprint.

Cite this