Date of Award

2011

Document Type

Doctoral Thesis

Degree Name

Doctor of Philosophy

Department

Electronic Engineering

First Advisor

Dr. Susan Rea

Second Advisor

Dr. Dirk Pesch

Abstract

Wireless Sensor Networks (WSNs) are a key enabling technology for industrial monitoring applications where the use of wireless infrastructure allows high adaptivity and low cost in terms of installation and retrofitting. Although WSNs present several advantages over traditional wired monitoring systems, there is significant concern as to the reliability of wireless communications due to the unpredictability of the wireless channel, which is inhibiting the widespread adoption of the technology. Furthermore, due to the diverse range of monitoring applications that are found in industrial environments and their differing requirements, there is a need for quality of service (QoS) provisioning so that the technology can be successfully adopted. However, achieving QoS provision in W'SNs is not a straightforward task due to the strict energy and memory limitations within sensor nodes as well as the bandwidth restrictions imposed by the technology. Fundamental QoS characteristics for industrial monitoring applications are bandwidth availability for bursty, event-based or data logging traffic, reliability, soft-delay bounds, data-relevance differentiation and network lifetime.

This thesis presents an algorithmic based QoS Framework for WSNs targeting monitoring applications in industrial environments. The proposed framework is composed of a suite of schemes located at the networking and MAC layers that work in conjunction with the lower layers, as defined by the IEEE 802.15.4 standard, to satisfy the QoS requirements of industrial monitoring applications while taking into consideration the limitations imposed by the resource constraints of the sensor nodes.

The performance of the proposed QoS framework is gauged against existing Standards and networking protocols for WSNs in industrial environments using a realistic computer simulation environment. The evaluation results highlight the limitations of existing approaches in fulfilling industrial monitoring application QoS requirements across a range of simulation environments and settings. In contrast, the proposed QoS Aware Framework, which simultaneously considers sensor nodes limitations and application QoS requirements, has been shown to satisfy application needs with low memory and computational overhead. For instance, the proposed lnRout route selection algorithm achieves gains from 4% to 60% in the number of successfully delivered packets and the proposed D-SeDGAM algorithm for bandwidth allocation achieves gains of 100% to 145% in terms of the number of packets delivered on time at the sink.

Access Level

info:eu-repo/semantics/openAccess

Share

COinS