On Long-Range Dependence in High-Bitrate Wireless Residual Channels

Wireless applications, protocols and simulators can significantly benefit from thorough understanding and accurate modeling of the medium access control (MAC) layer residual bit-errors at high bitrates. In this paper, we analyze and model the biterrors encountered at the highest achievable 11 Mbps data rate of an 802.11b wireless local area network. We employ autocorrelation, aggregation and variancetime analyses to reveal long-range dependence in the bit-error process. This renders traditional models (e.g., simple Markov, Poisson) ineffective and therefore we employ a multifractal wavelet model (MWM) to accurately characterize the bit-error process. Our results outline that the MWM captures the longrange-dependent behavior of the bit-errors quite accurately. We illustrate that the MWM outperforms high-order Markov chains in both channel modeling and complexity.