Avoiding shared resource conflicts in mobile sensor networks with multiple missions

Abstract

As sensor networks become increasingly autonomous and grow to include mobility and actuation, the need for predictability in the execution of complex missions becomes crucial. In this perspective, we propose a discrete event controller (DEC) as an effective framework for modelling task constraints, priority orders and task schedules of mobile sensor networks (MSNs) in charge of executing multiple missions. The focus is to show that, for such systems, risks of deadlock arise and that the proposed DEC is able easily to implement effective deadlock avoidance algorithms. Several experiments of coordination policies on an actual MSN test bed with shared heterogeneous resources are presented.