Abstract - lic thesis Jonas Neander, Ide, Mälardalens
Title: Using existing infrastructure as support for wireless sensor
Recent advancements in electronic design, such as low-power circuits,
efficient wireless communication, and improved energy supply, has
vision of wireless sensor networks to become a reality. Wireless sensor
typically consist of hundreds up to thousands of collaborating low-cost,
battery-driven and wireless sensor nodes with scarce resources. The
sensor nodes are typical small physical entities, and usually small as a
matchbox but can in extreme cases be no larger than a cubic millimeter.
In this thesis we present an architecture called AROS that uses existing
infrastructure to aid in the management of wireless sensor networks. As
example, the existing infrastructure could be situated in hospitals or
industrial buildings. The existing infrastructure can aid in prolonging
lifetime of the wireless sensor network by having ``unlimited'' energy,
range radio capacity, and high-speed computers. We enable prolonged
centralizing some of the energy consuming administrative functionality
wireless sensor networks.
We show, by simulations, that the AROS architecture is able to prolong
lifetime of the sensor nodes. AROS is compared to a well known cluster
architecture, LEACH. The comparisons show that AROS with static
performs at least as well as LEACH in small wireless sensor networks in
100x100m, and up to 97 % better in long distance wireless sensor
networks in the
size of 400x400m. We show that AROS still has got 88 % of its sensor
when LEACHs' network demises.
In our simulations we have also studied how dynamic network clustering
using a TDMA scheduler and non-mobile wireless sensor nodes, affects
of data received by a base station. We show that AROS is better than
collecting data to the base station with the same total amount of
long distance networks and that AROS performs as well or better than
small wireless sensor networks.