Skip to main navigation Skip to search Skip to main content

Eliminating explicit contention from channel access in distributed wireless networks

  • Aijing Li
  • , Chao Dong
  • , Fan Wu
  • , Shaojie Tang
  • , Guihai Chen
  • , Qihui Wu
  • Shanghai Jiao Tong University
  • Peoples Liberation Army Engineering University
  • Nanjing University of Aeronautics and Astronautics

Research output: Contribution to journal › Article › peer-review

Abstract

Emerging techniques such as Wi-Fi direct makes the objective of always-on be true. People can easily chat and share files with nearby friends even without AP (Access Point) or cellular coverage. In this paper, we focus on the channel efficiency issue of AP-free Wi-Fi networks, which can be easily constructed in the subway, in a high-speed railway, or when camping in the wild. Today IEEE 802.11 DCF is the most commonly used MAC protocol for Wi-Fi networks, however, due to the control messages and backoff time, channel efficiency in high data rate networks can be extremely low. To solve this, we propose CD-MAC, which allows control messages to be transmitted with data packets concurrently, and thus eliminates the overheads of backoff and explicit contention. To maintain the reception reliability, we redesign the control messages and use signal detection in PHY instead of bits decoding to detect them. In MAC layer, CD-MAC is built upon our Correlation Detection based PHY. We have implemented and evaluated CD-MAC using USRP N210. Evaluation results show that CD-MAC can achieve over 95.5% channel efficiency and provide throughput gains of up to 80%, 60%, and 29.1% compared with DCF, 802.11ec, and back2F, respectively.

Original languageEnglish
Article number8734182
Pages (from-to)70-82
Number of pages13
JournalChina Communications
Volume16
Issue number5
DOIs
StatePublished - May 2019

Keywords

  • ad hoc networks
  • correlation detection
  • implicit contention
  • media access
  • OFDM
  • USRP

Fingerprint

Dive into the research topics of 'Eliminating explicit contention from channel access in distributed wireless networks'. Together they form a unique fingerprint.

Cite this