Continuous approximation of stochastic models for wireless sensor networks

Stochastic analysis of wireless sensor networks becomes exceedingly hard as the number of nodes in a network grows large. In this paper we intend to address this issue by proposing a method of modeling large networks by dynamical systems rather than explicit Markov models, called Mean-Field Approximation. We verify the suitability of Mean-Field Approximation by analyzing ALOHA, by both studying a discrete model and a system of differential equations and then by comparing these models. We then extend the modeling technique in order to express characteristics of a network running a CSMA/CA protocol.