Virtual Sensor Networks: An Embedded Agent Approach
Richard Tynan, Gregory M. P. O’Hare, Michael J. O’Grady, Conor Muldoon · 2008
As the cost and size of WSN nodes reduce, it will be possible for deployments to provide node densities far in excess of the resolution requirements of any one single application. When this is the case, multiple, independent Virtual Sensor Networks (VSNs) can be maintained on the same physical deployment that satisfy distinct application requirements. In addition to this, nodes not part of any VSN can be considered redundant and can therefore be opportunistically hibernated. In this paper, the disadvantages of a layered approach to achieving such a system are demonstrated. An architecture based on a Multi-Agent System (MAS) is proposed as an alternative to alleviate some of the problems with the layered technique. Utilising a MAS facilitates the distribution of new VSNs onto a deployed WSN and permits heterogeneous timing regimes to be implemented that allow optimal performance of the individual VSNs.