Geometric networks based on Markov point processes with applications to mobile sensor networks
(With David Irons)
Preprint.
Abstract: This paper presents a Markov point process model for generating geometric graphs, capable of evolving over time and responding to external effects. Networks from these models are capable of self organising to appropriately cover a space despite a variety of obstacles or geographical constraints. These models have direct applications to mobile sensor networks, made up of mobile computational units capable of local environmental sensing and communication. In line with the desirable properties and deployment challenges for these systems, we evaluate summary statistics for our models; such as coverage, connectivity and robustness. We also demonstrate and evaluate how our models are applicable on three dimensional geographical terrains.
