Agents based control and coordination model for a distributed wireless sensor network to maximize field life.
Abstract
In recent years, wireless sensor network deployments for real life applications have rapidly
increased. However, energy control remains one of the major barriers hindering complete
exploitation of this technology because sensor nodes are usually powered by batteries with a limited
lifetime. Therefore efficient energy management is key requirement for the success of a wireless
sensor network. In this project, we proposed a distributed agent based model to solve the issue of
energy management in WSNs. In our approach, sensor nodes are modeled as agents. One of the
agents is a controller agent called the expert sensor agent (ESA). The ESA at the start of sensing
sets a threshold energy level. During sensing, the ESA continuously monitors each of the other
sensor agents’ energy levels and if a sensor node’s power level falls below the threshold, the ESA
deactivates the node and at the same time activates one of the nodes currently in sleep mode. At the
of the day, the sensing work load is equitably distributed across the nodes. The model was
implemented using simulation experiments in OPNET. We evaluated our agent based model against
none agent based technique in terms of power consumed and end to end delay (ETE). Our
simulation results indicate that our agent based model achieves over 75% power savings compared
to the non-agent techniques while at the same time ensuring real time delivery of data.
Citation
MASTERS OF SCIENCE IN COMPUTER SCIENCEPublisher
University of Nairobi School of Computing and Informatics