Markov and multifractal wavelet models for wireless MAC-to-MAC channels

Statistical understanding of bit-errors above the physical layer facilitates the design and verification of wireless protocols and applications. In this paper, we provide analysis and modeling of bit-errors at the 802.11b MAC layer. We show that MAC-to-MAC bit-errors at 2 and 5.5 Mbps are Markov, whereas those at 11 Mbps are long-range dependent (LRD). For the 2 and 5.5 Mbps channels, we observe that low-complexity hierarchical and hidden Markov models cannot characterize the bit-error processes, and consequently high-order full-state Markov chains are employed for accurate channel characterization. A multifractal wavelet model is employed to accurately capture the LRD 11 Mbps bit-error process.