An ensemble approach to analyse mobile sensor networks

Abstract

Mobile sensor networks (MSNs) are characteristic of dynamically changing topology due to node mobility. The traditional least overhead routing approach (LORA) to simulate MSNs cannot be applied to quantitatively assess measures of complex network analysis (like assortativity): for example, the assortativity index of one network snapshot need not be the assortativity index of another network snapshot. In this paper, we propose an 'ensemble' approach to collect the underlying information (like the degrees of the end vertices of the links) of the network snapshots over a period of time, and assess the targeted analytical measure (like assortativity index) of the MSN on this ensemble data that will be a representative of the topological changes over the entire simulation time period. The ensemble approach abstracts the node and edge ids and uses only the actual node-level metric values as the basis to calculate the network-level metric of interest.