Multi criteria adaptation mechanisms in homological sensor networks

Abstract

One fundamental feature of sensor network is unpredictable and this brings us a challenge of how to design an adaptation mechanism to let a sensor network adjust itself based on current "context" and pre-defined network criteria which include coverage density, fault tolerance, system life and so on. The key idea of our work is, based on calculation of nodes coverage, we can let networks be adaptive to ongoing changing network situation. We use system working life as our criterion in this work to demonstrate the potential usage of our system. An algebraic topology tool, homology, is used in our work to calculate sensor nodes coverage since the local properties of a sensor network, obtained by local interactions among nodes, can be captured by certain topological spaces. Also the global properties of the sensor network characteristics correspond to certain topological invariants of these spaces. Unlike other approaches, our work does not need any node location or orientation information and it does not have any assumption about the node deployment control and domain geometry. The only thing that it needs to calculate nodes coverage is a peer to peer communication graph.