Constant-Complexity Models for Wireless Channels

— High-order full-state Markov (FSM) chains have been employed to model errors and losses in many wireless studies. The complexity of this modeling paradigm is an exponential function of a random process’ memory-length and hence the viability of FSM chains in resource-constrained wireless environments is severely limited. In this paper, we address an important yet unsettled question: What characteristics of high-order FSM chains should be captured by a low-complexity approximate model? We analytically derive vital guidelines for accurate approximation of an FSM chain of arbitrary memory-length. These guidelines lead to a novel constant-complexity model (CCM), which always comprises of five states irrespective of a process’ memory-length. Our trace-driven evaluations for 802.11b and GSM channels demonstrate that the 5 state CCM, while providing orders of magnitude reduction in complexity, is comparable to FSM chains and better than linear-complexity models.

• Type: Conference
• Year: 2006
• Where: INFOCOM
• Authors: Syed A. Khayam, Hayder Radha
• Keywords: Communications, Fsm Chains, INFOCOM 2006, Low-complexity Approximate Model, Resource-constrained Wireless Environments