Sensoric CSP Networks

Specialeforsvar ved Martin Ellebæk og Rene Ravn


Demand and development of sensor networks is steadily increasing and is already being applied by many industries and researches worldwide. But collection and post processing of information in such sensor networks, traditionally exploits complex and inflexible units. Thus domain scientists with limited funding, find it hard to develop customized sensor networks to suit their specific situation.

Using Communication Sequential Processes (CSP) to break down complex sensor network units into smaller components generates an equivalent network focusing on simplicity, dynamism and stability. Components will be dynamically registered as a part of the autonomous network using automated negotiation techniques, thus easing configurations and verification of the network.

A complete environment for development and test of sensor components and networks will likewise increase the availability of sensor networks and ease the application. Focus of the individual researcher can thus be moved from the development of a network to the development of a particular functionality.

This thesis performs an initial analysis of the benefits arising from splitting of sensor components into smaller and simpler components. The analysis has subsequently formed the basis for implementing a prototype of a complete implementation and simulation tool for sensor networks. This tool will remove much of the complexity that normally characterizes the development of sensor networks, and thereby rendering domain scientists better capable of utilizing sensor networks.

Censor: Philippe Bonnet, ITU

Supervisor: Brian Vinter, DIKU