Distributed group management in sensor networks: Algorithms and applications to localization and tracking

The tradeoff between performance and scalability is a fundamental issue in distributed sensor networks. In this paper, we propose a novel scheme to efficiently organize and utilize network resources for target localization. Motivated by the essential role of geographic proximity in sensing, sensors are organized into geographically local collaborative groups. In a target tracking context, we present a dynamic group management method to initiate and maintain multiple tracks in a distributed manner, Collaborative groups are formed, each responsible for tracking a single target. The sensor nodes within a group coordinate their behavior using geographically-limited message passing. Mechanisms such as these for managing local collaborations are essential building blocks for scalable sensor network applications.