Energy Management in Sensor Networks

Research Overview/Summary of Contributions:
In the last few years, my co-workers and I have investigated the questions of optimal sleep scheduling and routing that can maximize the overall long-term utility of an energy-constrained sensor network. On the question of optimal node activation (sensor sleep scheduling) for rechargeable sensor networks, we have obtained activation policies that maximize the coverage utility of the network provably near-optimally for certain important classes of sensor network topologies. The problem of optimal sleep scheduling is fundamental question in sensor networking, and the uniqueness of our work lies in the consideration of rechargeable sensor systems, and in formalizing/addressing this question from a stochastic optimization perspective. Our key research results in this context include proving that spatial correlations degrade system performance, and developing a threshold policy that optimizes network utility irrespective of the degree of spatial correlations in the system [1,2]. Our contribution was immediately recognized as our initial paper on this topic was adjudged Best Paper Award Runner-Up in the prestigious IEEE Infocom 2005 [1] . Subsequently, we have investigated the question of optimal node activation (sleep scheduling) for a single sensor with the goal of maximizing the detection probability of temporally correlated events [3], and derived activation policies that are provably optimal or near-optimal, under different scenarios. The problems we have considered are fundanmental questions in the analysis and optimization of sensor networks, and our results significantly enhance the understanding of how sensor networks should be deployed and managed. Some of these algorithms have been implemented in a small sensor network prototype; a larger scale deployment is planned in Lake George, NY.

My research on this topic is supported by NSF.

The key research papers on this topic are:

1. K. Kar, A. Krishnamurthy and N. Jaggi, Dynamic Node Activation in Networks of Rechargeable Sensors, IEEE/ACM Transactions in Networking, Vol. 14, No. 1, pp. 15-26, February 2006.
2. N. Jaggi, K. Kar and A. Krishnamurthy, Near-optimal Activation Policies in Rechargeable Sensor Networks under Spatial Correlations, Accepted for publication in ACM Transactions on Sensor Networks, September 2007.
3. N. Jaggi, K. Kar and A. Krishnamurthy, Rechargeable Sensor Activation under Temporal Correlations, Accepted for publication in Wireless Networks, September 2007.

Other papers on this topic can be found on my full publication page.