Energy efficient and load-balanced routing schemes for in-network data aggregation in wireless sensor networks / Omar Adil Mahdi

• Main Author: Omar Adil , Mahdi
• Format: Thesis
• Published: 2017
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://studentsrepo.um.edu.my/12024/1/Omar_Adil.pdf; http://studentsrepo.um.edu.my/12024/

Wireless sensor networks (WSNs) constitute a group of small autonomous units known as sensor nodes within an infrastructure-less and self-configuring wireless network. In such networks, the deployment of high-density sensor nodes creates a possibility for close nodes to produce redundant data that leads to unnecessary energy consumption. Therefore, efficient energy management strategies that reduce data redundancy have been proposed by prior researchers. This includes the well-known Data Routing for In-Network Aggregation (DRINA) and Efficient Data Collection Aware of Spatio-Temporal Correlation (EAST) that utilize in-network data aggregation to reduce data redundancy. The two protocols maximize the formation of overlapping paths to increase aggregation rate. Their methods, however, do not consider the actual nodes energy resources and the traffic load that could result in early multi-hop relay nodes failure. This affects the network ability to forward packet successfully, thus, lead to unstable network structure. In this thesis, energy-efficient and load-balanced routing schemes are proposed. The proposed schemes extend the work in DRINA and EAST protocols. The first scheme is the Weighted Data Aggregation Routing Scheme (WDARS) that aims to maintain an acceptable trade-off between energy efficiency and routes overlapping by considering metrics related to energy efficiency and load balancing for multi-hop relay nodes and routes overlapping point selection. The second scheme is the Weight Aware Spatial-Temporal Correlation Routing Scheme (WST-RS) which exploits the spatial and temporal correlation among the sensor observations in a fully distributed manner. This thesis also introduced a new Multi-Criteria Node Weight (MCNW) metric to weight the node’s status in terms of various metrics that contributes as a link cost function for link quality assessment and selecting the lightweight routes to the destination depending on the status of nodes. The results show that WDARS scheme outperforms DRINA and InFRA in terms of energy consumption by 11.65% from the initial energy as compared to 21.75% and 35.71% respectively. Similarly, the nodes in the proposed WST-RS scheme records the lowest energy consumption with 16.25% from the initial energy compared to the DRINA (35.79%), YEAST-CF (24.83%), and EAST (20.58%). The first node to die in WST-RS is much longer, nearly 3.06, 2.22, 1.82 times longer than DRINA, YEAST-CF, EAST. This research will contribute toward effectively alleviating the energy constraints associated with WSNs by extending the network lifetime and increasing the availability of services even in highly dense networks with heavy traffic load.