[Truncated abstract] Developing wireless sensor network communication protocols for industrial environments is a challenging task. The wireless channel conditions in industrial environments are harsher as a result of multipath propagation of radio signals within an environment where the mechanics of the surrounding industrial activities often lead to severe small-scale fading effects. The design of network protocols to function in such an environment needs to provide a robust communication platform for the wireless sensor nodes, while optimizing the utilization of the limited node resources available. As existing general MAC protocols for wireless sensor networks do not work well under harsh channel conditions, we address this problem in this thesis by proposing a medium access control (MAC) protocol framework for wireless sensor networks in industrial environments. To describe the impact of an automated industrial site on small-scale fading effects in an industrial indoor wireless network, we use a site-specific ray-tracer for predicting signal propagation based on building blue prints to simulate the signal propagation paths through an industrial site with periodically moving objects. We found that in a fully automated industrial site, the periodic movements of objects with constant velocities result in an approximately periodic distribution of fading periods in the channel. Based on this finding, we propose a link state dependent TDMA-based MAC protocol for wireless sensor networks designed for automated industrial applications. ... This technique also conserves energy and maximizes packet integrity as the sensor nodes avoid performing network activities when the channel is sampled to be affected by fading. As a measure for fault-tolerance, we also propose a dynamic link reconstruction technique that allows sensor nodes to reconstruct new parent links when their present links are severely affected by fading. MAC protocols that adopt active buffering in fading-affected channels suffer bufferoverflow and latency issues as a result of the postponement of scheduled transmissions during periods when the channel is in a fade. Consequently, time-sensitive data packets that may contain critical information may miss their deadlines in a severely affected fading channel. Hence, we propose a general fading-aware data management (FADE) MAC protocol extension that uses buffer nodes to offload the memory buffer off sensor nodes in the network and prioritizes traffic based on a simple proposed priority scheme. The FADE extension provides a balance balance between minimum end-to-end latency for critical event reporting, high packet delivery guarantee, low energy consumption, and minimum buffer requirements on the sensor nodes. In summary, this thesis presents the design, implementation, and evaluation of all our proposed network protocols that are combined as a framework for wireless sensor networks in industrial environments.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - 2009|