Sensor networks present many opportunities for changing the way computers interact with the environment. Sensor network devices present major challenges because they must be functional under strict power requirements. Since wireless communication is the largest source of power consumption, we explore efficient algorithms for information gathering that try to minimize the amount of communication while still gathering the requisite amount of data.
Our research focuses on an efficient traversal method to gather information from the network. We assume that nodes have a fixed communication radius. Our model also assumes that every node has complete information about each of its neighbors. In addition, a node can go into sleep mode, and it will not take part in communication while in this state. Given the above constraints, we are attempting to design heuristic methods that will gather information from all nodes while actually visiting a minimal number of nodes.
Our approach involves a graph theoretical formulation coupled with geographic knowledge about the network. In this approach, each node is represented as a vertex and a communication path is represented as an edge. We are investigating the significance of network perimeter nodes, articulation points, and sub-areas as components of this heuristic approach.