Edinburgh Napier University

  Sensor nodes and their underlying communication technologies are characterised by restricted power resources, restricted processing, limited storage capacities, low data rates and lossy links, and they may also comprise up to a thousand nodes. Wireless Sensor Networks (WSNs) require effective methods for data aggregation, forwarding and processing in order to preserve the limited nodes resource.

Energy efficiency in WSNs has been widely investigated; it is still a challenging dilemma, and new mechanisms are required to fulfil the identified gaps in the literature. In most of WSNs applications, the energy cost has a significant effect on the network lifetime. Along with the energy efficiency, the Network Scalability and Data Reliability are other challenges affecting the performance of robust sensor network. Consequently, network reliability increases at the expense of energy consumption due to the traffic generated to maintain forwarding paths. On the other hand, scalability is a key component of WSNs because, for example, the network may need to grow to cover more space. Therefore, adding more nodes to the network will increase the number of data traffic within the network. Thus, these challenges impose the need to develop new mechanisms that cope with this sensor network requirements and be able to scale while providing efficient data routing and less energy consumption.

Clustering mechanisms are among the most commonly recommended approaches by the research community to sustaining a sensor network throughout its lifetime and provide a scalable architecture and reliable data delivery. Despite a number of research activities associated with clustering in WSNs, some aspects of clustering have not yet been adequately investigated. The conflict between high energy consumption and reliability results in excessive energy waste; this is mainly caused by a high number of control messages exchanged to select cluster-head nodes and frequent re-clustering process in traditional cluster-based mechanisms. The re-clustering of the entire network in each round, or when one of the clusters depletes their energy, is not effective because of the extensive overhead and re-clustering process used.

The former problems have been solved using tree novel mechanisms that enhance the network's Energy-Efficiency, Reliability and Scalability while considering the node's limited resources. The performance of the proposed mechanisms is validated under realistic network settings through extensive simulation experiments under different scenarios. The realistic energy consumption model is considered based on Chipcon CC2420 advanced radio modules implemented in Castalia simulator. The results obtained revealed that proposed mechanisms outperform the existing mechanisms in term of energy consumption while achieving reliability and scalability.