Please use this identifier to cite or link to this item:
|Title:||Distributed autonomous systems to track mobile nodes in industrial wireless sensor networks|
|Abstract:||With an ever-increasing trend in automating industrial processes and systems, and with a need to collect and manage information and sensory data in an economic, secure, reliable, and real-time fashion, wireless sensor networks have arisen as one promising solution. One of the imperative problems in the realm of wireless sensor networks is the problem of wireless sensors localization. Although much research has been conducted in this area, many of the proposed approaches produce unsatisfactory results when exposed to dynamic and noisy conditions of a manufacturing environment. This research, first, presents a robust solution, based on artificial neural networks, to tackle the problem of mobile node tracking in industrial wireless sensor networks. The proposed technique is tested on a simulation model, and the results obtained are validated with a physical small-scale wireless sensor network. A sensitivity analysis follows the simulation and physical model studies to investigate the impact of a set of ambient factors and wireless sensor network parameters on the proposed technique. Much of the studies in the literature of the mobile node tracking merely focus on the accuracy of the localization technique. Nevertheless, to design a successful tracking technique for industrial wireless sensor networks requires one to address many more concerns. In industrial wireless sensor networks, where hundreds to thousands of wireless nodes are deployed, resource management, response time, load distribution, reconfigurability, and adaptability of a tracking system become of paramount importance. Thus, this research presents three tracking systems to position mobile nodes in industrial wireless sensor networks in a soft real-time fashion. Each tracking system employs a unique wireless sensor network topology, namely static clusters, dynamic clusters and ad hoc networks of wireless nodes. Each tracking topology is, in turn, governed by a unique multi-agent system. Subsequently, a comprehensive experimentation based on statistical designs is performed to analyze and compare the performance and efficiency of the proposed tracking systems with respect to a set of formulated metrics. The results obtained illustrate that while ad hoc networks exhibit higher performance, the static and dynamic clusters excel in efficiency related metrics.|
|Appears in Collections:||Electronic Theses|
Files in This Item:
|UCalgary_2014_Gholami_Mohammad.pdf||3.03 MB||Adobe PDF||View/Open|
Items in The Vault are protected by copyright, with all rights reserved, unless otherwise indicated.