Performance Comparison of Clustering Schemes in Sensor Networks

Recent advances in microelectro-mechanical systems and wireless communications have enabled the development of distributed, wireless networks of small, inexpensive, low power sensors. These sensors, which possess sensing, data processing and short range communication capabilities, can be deployed in diverse environments to collect useful information. Large numbers of such inexpensive sensors form ad hoc wireless networks which can be used in a variety of commercial and military applications.

As sensors become inexpensive and widely available, sensor networks can be deployed in larger scales. One of the issues arising consequently is scalability. A flat structure of a large number of sensors often provides low scalability and makes network-wide coordination difficult. To solve this problem, hierarchical architectures (clusters) have been proposed to solve the scalability problem. Appropriate clustering can reduce the need for global coordination and restrict most of the sensing, data processing and communication activities within clusters, thus can improve resource effciency and prolong network lifetime. Clustering can also provide load balancing if appropriately configured.