M. Naznin, X. Du, and K.E. Nygard (USA)
Wireless sensor network, collaborative strategies, decision making, heuristic, collaborative systems, co operative nodes.
Obtaining a high level of coverage is a major challenge in sensor networks, due to the short battery life time of sensor nodes and the complexity of cooperatively controlling large numbers of sensors. We address the fundamental problem of scheduling wake and sleep times for multiple sensors to ensure adequate coverage of a geographical area over time with low energy expenditure. Most of the work on coverage algorithms makes restrictive assumptions, such as requiring uniform coverage over the entire area, homogeneous sensors, continuously active sensors, and simplistic models of sensor performance. Our approach incorporates more realistic models of sensor performance that incorporate probabilities of detection when multiple sensors are simultaneously employed, and the concept of exposure of an invader to sensor detection over time as it follows a path. Required coverage levels for points of interest are explicitly differentiated. We also incorporate the concept of variability in the alertness requirements that the sensor network is expected to maintain. Times of high alertness assign large numbers of sensors to be actively sensing most of the time, with correspondingly high detection probabilities. Times of lower alertness are supported with small numbers of sensors that are scheduled to periodically go into a sleep state to conserve energy. The method is energy efficient and load-leveling since it tries to minimize the maximum deviation from the number of sensors and active times required to maintain the prescribed alertness.
Important Links:
Go Back
