Loss performance model for wireless channels with autocorrelated arrivals and losses

  • Authors:
  • D. Moltchanov;Y. Koucheryavy;J. Harju

  • Affiliations:
  • Institute of Communications Engineering, Tampere University of Technology, P.O. Box 553, Tampere, Finland;Institute of Communications Engineering, Tampere University of Technology, P.O. Box 553, Tampere, Finland;Institute of Communications Engineering, Tampere University of Technology, P.O. Box 553, Tampere, Finland

  • Venue:
  • Computer Communications
  • Year:
  • 2006

Quantified Score

Hi-index 0.24

Visualization

Abstract

In this paper, we firstly propose simple and computationally efficient wireless channel modeling algorithm that explicitly takes into account first- and second-order statistics of frame error observations. For this purpose we use a discrete-time Markov modulated process with at most single event (error) at any time slot. Then, we identify a class of priority queuing systems of G+G/GI/1/K type capable to model the frame transmission process over wireless channels with correlated arrival and loss processes. Using the proposed frame error process, performance evaluation model of the wireless channel at the data-link layer is then reduced to the special case of @?"iD-BMAP"i/D/1/K queuing system with non-preemptive priority discipline. The proposed queuing representation allows to capture forward error correlation (FEC) and automatic repeat request (ARQ) functionality of the data-link layer as well as distributional and autocorrelational properties of the frame arrival and frame loss processes. This model is further analyzed for a number of performance parameters of interest including probability function of the number of frames in the system and probability function of the number of lost frames. It is shown that the channel response in terms of the mean number of frames in the system and the mean number of lost frames varies substantially for different first- and second-order frame error and arrival statistics. The impact of input statistics is also different for normal (@r=1) of @?"iD-BMAP"i/D/1/K queuing system.