Energy-Efficient Dynamic Spatial Resolution Control for Wireless Sensor Clusters

Abstract

In very large-scale dense sensor network applications, more sensor nodes may be deployed than are required to provide the initial desired spatial resolution. Such over-deployment can extend network life, improve robustness, and accommodate network dynamic. To enable large deployments, tiered and clustered network structures may be adopted for scalability and manageability. This article presents a highly scalable, distributed control method to manage the activity of sensors in each cluster so that dynamic application-specific spatial resolution requirement can be achieved with minimum energy cost and with uniform participation of sensors. The method, Look-Ahead Resolution Control (LARC), utilizes a look-ahead prediction of activities of sensors in the cluster and provides feedback to nodes as part of acknowledgments to transmissions. LARC is shown to be highly responsive to system dynamics such as changes in the resolution requested, node failures or replenishment. Existing control methods not only fall short in network life and scalability, but also do not provide such responsiveness and uniform participation to ensure a diverse representation of sensed data. LARC is extended to tradeoff between energy and delay in recovery from failures of large number of nodes, by multiplexing sleeping, listening, and transmissions probabilistically in such a way that the control overhead is minimized while the delay is bounded. The article presents the control strategy for LARC along with performance studies showing near-uniform participation and near-optimal cluster lifetime.

Keywords:

• Large-Scale Sensor Networks
• Spatial Sensing Resolution
• Distributed Control

This research was supported in part by Vermont NSF EPSCoR grant EPS 0236976. Also, support of the Vermont Advanced Computing Center under a NASA grant is gratefully acknowledged.

Notes

Part of this work was published in Proceedings of the 1st Workshop on Resource Provisioning and Management in Sensor Networks (RPMSN ‘05) in Conjunction with the 2nd IEEE International Conference on Mobile Ad Hoc and Sensor Systems, Washington, DC, Nov. 2005.