Authors: Srikanth Sundaresan, Israel Koren, Zahava Koren, C. Mani Krishna
Addresses: College of Computing, Georgia Institute of Technology, Atlanta, GA, USA. ' Electrical and Computer Engineering, University of Massachusetts, Amherst, MA, USA. ' Electrical and Computer Engineering, University of Massachusetts, Amherst, MA, USA. ' Electrical and Computer Engineering, University of Massachusetts, Amherst, MA, USA
Abstract: Energy conservation is a major issue in the management of sensor networks. In this paper, we consider the problem of optimising the performance of energy-constrained sensor networks in the context of event-detection applications, with the performance metric being a weighted average of the numbers of missed events and false alarms. We propose two adaptive event-driven techniques, both simple to implement at the node level, that change the nodes| duty cycles based on the patterns of detected events. Simulation results show that this approach is effective in both cutting down wasteful energy consumption as well as in improving the detection performance in applications characterised by long periods of quiescence followed by bursts of activity.
Keywords: sensor networks; environment monitoring; adaptive duty-cycling; MDP; Markov decision process; energy conservation; event clustering; energy consumption; simulation.
International Journal of Sensor Networks, 2009 Vol.6 No.2, pp.89 - 100
Received: 13 Jul 2008
Accepted: 02 Feb 2009
Published online: 26 Oct 2009 *