Distributed optimization in sensor networks

Wireless sensor networks are capable of collecting an enormous amount of data over space and time. Often, the ultimate objective is to derive an estimate of a parameter or function from these data. This paper investigates a general class of distributed algorithms for "in-network" data processing, eliminating the need to transmit raw data to a central point. This can provide significant reductions in the amount of communication and energy required to obtain an accurate estimate. The estimation problems we consider are expressed as the optimization of a cost function involving data from all sensor nodes. The distributed algorithms are based on an incremental optimization process. A parameter estimate is circulated through the network, and along the way each node makes a small adjustment to the estimate based on its local data. Applying results from the theory of incremental subgradient optimization, we show that for a broad class of estimation problems the distributed algorithms converge to within an /spl epsi/-ball around the globally optimal value. Furthermore, bounds on the number incremental steps required for a particular level of accuracy provide insight into the trade-off between estimation performance and communication overhead. In many realistic scenarios, the distributed algorithms are much more efficient, in terms of energy and communications, than centralized estimation schemes. The theory is verified through simulated applications in robust estimation, source localization, cluster analysis and density estimation.

[1]  Kung Yao,et al.  Source localization and beamforming , 2002, IEEE Signal Process. Mag..

[2]  Saurabh Ganeriwal,et al.  Aggregation in sensor networks: an energy-accuracy trade-off , 2003, Ad Hoc Networks.

[3]  Leonidas J. Guibas,et al.  A Distributed Algorithm for Managing Multi-target Identities in Wireless Ad-hoc Sensor Networks , 2003, IPSN.

[4]  Keith D. Kastella,et al.  Sensor Management Using Relevance Feedback Learning , 2003 .

[5]  S. Ganeriwal,et al.  Aggregation in sensor networks: an energy-accuracy trade-off , 2003, Proceedings of the First IEEE International Workshop on Sensor Network Protocols and Applications, 2003..

[6]  Benjamin Van Roy,et al.  Distributed Optimization in Adaptive Networks , 2003, NIPS.

[7]  Michael I. Jordan,et al.  On Convergence Properties of the EM Algorithm for Gaussian Mixtures , 1996, Neural Computation.

[8]  D. Bertsekas,et al.  Incremental subgradient methods for nondifferentiable optimization , 1999, Proceedings of the 38th IEEE Conference on Decision and Control (Cat. No.99CH36304).

[9]  Akbar M. Sayeed,et al.  Collaborative Signal Processing for Distributed Classification in Sensor Networks , 2003, IPSN.

[10]  B. Ripley,et al.  Robust Statistics , 2018, Encyclopedia of Mathematical Geosciences.

[11]  D. Bertsekas,et al.  Convergen e Rate of In remental Subgradient Algorithms , 2000 .

[12]  Robert D. Nowak,et al.  Distributed EM algorithms for density estimation and clustering in sensor networks , 2003, IEEE Trans. Signal Process..

[13]  Deborah Estrin,et al.  Scalable Coordination in Sensor Networks , 1999, MobiCom 1999.

[14]  Dimitri P. Bertsekas,et al.  Incremental Subgradient Methods for Nondifferentiable Optimization , 2001, SIAM J. Optim..

[15]  Yu Hen Hu,et al.  Energy Based Acoustic Source Localization , 2003, IPSN.

[16]  Kannan Ramchandran,et al.  On compression for robust estimation in sensor networks , 2003, IEEE International Symposium on Information Theory, 2003. Proceedings..