Exploiting Multi-Channel Clustering for Power Efficiency in Sensor Networks

Sensor networks typically comprise of a number of inexpensive small devices with processing, communication and sensing abilities that collaborate to perform a common task. Sensor devices use batteries as their sole power supply. The operational lifetime of a sensor network, therefore, depends entirely on the better utilization of the devices. Typically a sensor network is divided into clusters to optimize power utilization by performing division of labor and data aggregation within a cluster. This paper introduces a novel approach to naturally distributed clustering of sensor nodes in a sensor net using multi channel data planes. Our technique incorporates a virtual sense mechanism that reduces energy spent in sampling and transmission. It also decreases network traffic, thereby decreasing contention, potential collisions and retransmissions. This approach inherently implements a sleep-awake mechanism based on virtual sensing that contributes towards increasing the network lifetime by efficient utilization. The proposed technique can be used to track spreading phenomenon like forest fires and water flows. A spreading phenomenon can be represented by a field whose value changes dynamically with time over area. We focus on following the movement of such a dynamically changing field rather than obtaining the value of the field at different locations at disjoint random times