A separation principle for resource allocation in industrial wireless sensor networks

Abstract

Industrial production lines have been used to assemble a wide range of commercial goods such as metallurgy, automobile, and electronic devices. Since these production lines create tens of trillions of dollars annually, their production efficiency, cost, and safety are critical for global economy. This paper uses industrial wireless sensor networks (IWSNs) to monitor multi-stage production lines. Unlike traditional surveillance WSNs, IWSNs feature a unique cascaded network topology, which can be leveraged to optimize network performance (e.g., end-to-end delay). To our best knowledge, research along this direction is lacking. Specifically, considering the physical characteristics and functional requirements of production lines, we introduce the cascaded FieldNets where each FieldNet is a field sub-net corresponding to one process stage. In particular, the end-to-end minimization oriented resource allocation problem is concerned. It is a nonlinear mixed integer programming problem formulated by both (1) channel allocation among FieldNets and (2) multichannel transmission scheduling within each FieldNet. To solve it, a separation principle is proposed, by which we prove that the resource allocation within each FieldNet can be determined independently from the channels allocation among FieldNets. Performance evaluation demonstrates that the proposed resource allocation approach provides a $10{\times }$ larger region of schedulability and achieves as low as 10 % of end-to-end delay compared with the scheduling approach in WirelessHART, and only consumes half of the energy based on some existing MACs such as Y-MAC and EM-MAC under high-traffic condition.