Prediction of Channel Utilization with Artificial Neural Networks Model in Mac Layer in Wireless Local Area Networks

Karakurt H. B., KOÇAK C., ÖZKAN M. T.

WIRELESS PERSONAL COMMUNICATIONS, vol.126, no.4, pp.3389-3418, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 126 Issue: 4
  • Publication Date: 2022
  • Doi Number: 10.1007/s11277-022-09870-0
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, PASCAL, Applied Science & Technology Source, Communication Abstracts, Compendex, Computer & Applied Sciences, INSPEC
  • Page Numbers: pp.3389-3418
  • Keywords: Artificial neural networks (ANN), Channel usage, Receiver channel utilization%, Transmitter channel utilization%, Wireless local area networks, ALGORITHMS
  • Gazi University Affiliated: Yes


The use of channels in WLANs affects data communication. RTS Fragmentation Threshold (RTSFT), Fragmentation Threshold (FT), and Buffer Size (BS) input values; Receiver Channel Utilization% (RCU%) and Transmitter Channel Utilization% (TCU%) affect the output values. To use the communication channel most efficiently, RTSFT, FT, and BS input variable values, which give the best RCU% and TCU% values, were obtained using the Riverbed Modeler simulation tool. Bandwidth, which is an important parameter for increasing the QoS of Wireless Local Area Networks, provides only physical performance efficiency. The efficiency at which the bandwidth is used is extremely important for the pre-design of network topology. The channel utilization rate increases the efficient use of bandwidth in WLANs. RTSFT, FT, and BS, which affect the service quality in the MAC layer in WLANs, directly affect the quality of service. In this study, the Prediction of Receiver Channel Utilization% (RCU%) and Prediction of Transmitter Channel Utilization% (TCU%) performance criteria were predicted by using ANN algorithms. Using the Riverbed Modeler simulation tool, 125 different input variables with 11 nodes were selected and channel utilization was obtained. Used ANN data was obtained from real Riverbed Modeler simulation. Input variables (RTSFT, FT, and BS) against the output parameters (RCU% and TCU%) were obtained from ANN model prediction. It is concluded that the estimated RCU% and TCU% output values are revealed the affected directly proportional by the BS input and inversely proportional to the RTSFT and FT input values.