Enery-Efficient Network Management Protocols for Cognitive Radio Sensor Networks

Abstract

A cognitive radio sensor network (CRSN) is a wireless sensor network (WSN) in which the sensor nodes are equipped with cognitive radio. CRSN is envisioned as a key technology to support the future wireless networks such as seamless telecommunication, the Internet of things, and the improvement of spectrum utilization. Despite of its merits, CRSN yields numerous challenges: some being inherent to cognitive radio properties and others to WSN characteristics. Cognitive radio possesses dynamic spectrum access capability. Thus, it requires a spectrum decision method to select its operating channel. Furthermore, clustered topology is mostly favored in a WSN. However, cluster formation becomes challenging in a CRSN environment. Last but not least, a WSN as an application-driven network needs to maintain an acceptable degree of reliability, which is supported by the transport protocol. Above all, energy conservation is of the utmost importance because of the energy- and resource-constrained nature of sensor nodes. In this thesis, energy-efficient network management protocols for CRSNs are proposed, which are composed of a spectrum decision framework, a clustering protocol, and a transport protocol. The spectrum decision framework is distributed and it contains two spectrum selection algorithms: random selection and game-theory based selection. This framework also incorporates two spectrum sensing schemes of full and partial spectrum sensing, a simple clustering, a spectrum characterization scheme implementing Markov chain, and a cluster member coordination scheme. The clustering protocol is compact and it efficiently achieves compact cluster formation by adopting two sub-phases (cluster head discovery and cluster member invitation) of cluster formation. By introducing a novel concept of temporary support nodes, the clustering enables sensor nodes to form clusters efficiently. The transport protocol is a content-aware data transmission and acknowledgment method that aims to increase the network lifetime while decreasing delay and maintaining reliability at the same time. The performance of each proposed protocol is evaluated by means of computer simulations and compared with the existing works. The performance evaluation of the spectrum decision framework shows that the framework outperforms the existing work in terms of network lifetime and coordination overhead. The performance evaluation of the clustering protocol shows that it achieves outstanding energy savings that prolong the network lifetime and decreases both the clustering overhead and the average distance between cluster heads and their members, compared to the existing work. Finally, the performance evaluation of the transport protocol shows that it achieves remarkably longer network lifetime and shorter event-detection delay compared to the conventional transport protocol while preserving event-detection reliability.