On efficient clustering of wireless sensor networks

Wireless sensor networks are poised to increase the efficiency of many applications, such as target detection and disaster management. Typically sensors collect data about their surrounding and forward it to a command center, either directly or through a base-station (gateway). Due to inhospitable conditions, sensors may not always be uniformly deployed in the area of interest. Some sensors can be unreachable to the gateway due to their distance or the existence of obstacles in their communication path. Additionally, in many applications a large set of sensors is usually deployed and network scalability becomes a major concern. This paper investigates two different methodologies for clustering sensor networks with the goal of increasing the sensor coverage and availability allowing for dependable and efficient operation for large network setups. The first approach deploys multiple gateways and partitions sensors among these gateways. In the second approach some sensors are designated as agents for a single gateway in order to reach out-of-range nodes. The performance of both approaches is compared in a simulated environment.