Cellular Automata for Modeling Wireless Sensor Networks

Abstract

We propose a new approach for modeling wireless sensor networks based on the theory of cellular automata. Each sensor node is described by a cell, and the cell dynamics is governed by the transition rules or operators. It is shown that the corresponding transition function can be represented as a superposition of two functions, local and nonlocal. The local function is completely determined by the algorithm implemented in the microcontrollers of the nodes. The non-local function describes message exchange via radio channel. The proposed model allows us to transfer the algorithms from the model to microcontrollers of real sensor nodes. The developed approach is used to model a wireless sensor network operating in data acquisition mode. A simple algorithm is proposed for the network self-organization in tree-like structure with the base station as the root. The model of such a network is investigated and characteristics of its dynamics are studied, such as time of self-organization and data throughput. The proposed algorithm was also realized on real sensor nodes. We show that the real wireless network operates in agreement with the developed theory.