Sensor network architecture for multi-media traffic

The desire for more intelligence in the battlefield has given rise to the idea of routing video images over wireless sensor networks. This would apprise the combat decision makers with actual images of battlefield developments and allow them to make sound decisions. To achieve this objective, the characteristics of video traffic must be studied and understood. This paper focuses on evaluating the possibility of routing video images over a wireless sensor network. Video traffic is modeled and simulations are performed via the use of the Sun Small Programmable Object Technology (Sun SPOT) Java development kits configured in three different network topologies: the star topology, binary tree topology and chain topology. It is known that video traffic is self-similar and can be obtained by aggregating a large number of On-Off message sources. Hence, an On-Off model using Pareto distribution function is used to model video traffic over the network. The performance of each topology is evaluated based on parameters like mean throughput, mean interarrival time, mean packet drop, and mean delay. The results of the simulation reveal that each topology has its own merit.