Modelling and Performance Evaluation of Mobile Ad Hoc Networks

Abstract

Mobile ad hoc networks are characterized by having nodes that are self-organized and cooperative without any kind of infrastructure, being the most promising upgrade of the current telecommunication systems. The mobility and multihop capability of these networks allows the network topology to change rapidly and unpredictably, turning necessary the development of appropriate models to describe the multihop connectivity and the dynamic of multihop paths.

The research carried on in this dissertation starts by addressing the multihop connectivity for one-dimensional and two-dimensional ad hoc networks. The hop count probability distributions are derived when the underlying node spatial distribution is drawn from a Poisson process and, by using a Poisson randomization technique, when a fixed number of relay nodes are uniformly distributed in a region of interest. Numerical results illustrate the computation of the hop count probabilities.

We then present an analytical framework to characterize the random behavior of a multihop path by means of a piecewise deterministic Markov process. The mean path duration and the path persistence metrics are obtained as the unique solution of a system of integro-differential equations, and a recursive scheme for their computation is provided. Numerical results are presented to illustrate the computation of the metrics and to compare the associated results with independent link approximation results