Fault Tolerant Wireless Sensor Networks for Internet of Things
Loading...
Date
item.page.authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This research presents a comprehensive fault-tolerant framework for WSNs in IoT applications. First, a Vague Set-based fault diagnosis mechanism is proposed to accurately detect and classify sensor node failures. The classification system effectively distinguishes between fully operational nodes, partially faulty nodes, and dead nodes, facilitating optimal network management. Additionally, a faulty node reuse mechanism is introduced, which strategically repurposes partially faulty nodes to extend network lifetime and maintain connectivity, thereby reducing unnecessary energy wastage and premature network failure.
newlineTo enhance the accuracy and efficiency of fault detection, this study further proposes a two-population Z-test-based fault detection mechanism integrated with a majority voting scheme. This statistical approach ensures robust detection of heterogeneous hardware failures while minimizing false positives and false negatives, leading to a more energy-efficient fault management system. Furthermore, recognizing the challenges posed by intermittent connectivity in mobile and large-scale WSN deployments, this research employs Support Vector Regression (SVR)-based hop count prediction to optimize data forwarding. By leveraging machine learning techniques, the proposed scheme intelligently estimates hop count in intermittently connected networks, enabling efficient routing decisions that reduce latency and overhead while improving data delivery success rates.
newlineFinally, this thesis introduces an enhanced PRoPHET+ data routing protocol that incorporates cooperative network attributes for efficient communication in partially faulty networks. This routing mechanism considers various network parameters such as node buffer, bandwidth, power level, and deliverability probability to make intelligent forwarding decisions, thereby enhancing network performance in terms of reliability, energy consumption, and overall lifetime
newline