A novel location energy sleep mode saving algorithm (LESMS) based on clustering for WSN

Al-Baadani, Faris (2017) A novel location energy sleep mode saving algorithm (LESMS) based on clustering for WSN. Doctoral thesis, Anglia Ruskin University.

[img]
Preview
Text
Accepted Version
Available under the following license: Creative Commons Attribution Non-commercial No Derivatives.

Download (4MB) | Preview

Abstract

Due to sensors being distributed in an ad-hoc manner and their ability to sense different environmental conditions, Wireless Sensor Networks (WSNs) have become very popular in various real time applications and Internet of Things (IoT). The WSN is widely used in monitoring harsh environments and taking decisions on the basis of collected reports. Clustering in WSNs is a proven technique to avoid redundant transmissions to the sink. This allows for better utilisation of scarce network resources such as energy. Most of the clustering algorithms proposed in the literature involve a high number of message exchanges, which results in unnecessary energy consumption. This thesis proposes a new cluster head (CH) selection protocol based on the average energy and the location of the nodes. Five different approaches for the CH selection were examined and were statistically analysed: (i) random selection; (ii) selecting the node closest to the arithmetic mean node coordinate; (iii) selecting the node at the medoid coordinate; (iv) selecting the node nearest to the region centre; (v) selecting node nearest the BS. The mean distance of all nodes to their local CH was the dependent variable. Furthermore, three node grouping schemes were further studied and compared to the k-means clustering: (i) all nodes in the same group; (ii) dividing the sensor field into four rectangular quadrants and allocating nodes accordingly; (iii) dividing the sensor field into eight sectors and allocating nodes accordingly. T-Test and one way ANOVA were used for the P-value analysis. The sleeping mode technique was implemented. The residual energy level was then used as a main factor, for the CH selection. The average energy value and the minimum distance value both became the weighting factors to select the CH. The Assistant Cluster Head (ACH) technique was introduced to the LESMS to act as a backup system in case of the CH power levels reach critical levels or total loss of the CH node . Finally, a new technique was introduced to perform a periodic CH health check-up to monitors the energy level of the CH. Simulations were performed using MATLAB and Network Simulator 2 (NS2). Results show that selecting the node closest to the arithmetic mean as CH has outcome other approaches. Also, the sensor field dividing scheme has a shorter distance of all the nodes compared to the k-mean. The LESMS, under a mixture of workload environments resulted in low power consumption, low delay, low packet drop, and low control overhead. It also showed high throughput, high packet delivery ratio (PDR) and an increase in the total residual energy.

Item Type: Thesis (Doctoral)
Keywords: Wireless Sensor Networks, Internet of Things, Clustering, Cluster Head, Base Station, Location and Energy Sleep Mode Saving Algorithm, Arithmetic Mean, k-mean, Residual Energy, Assistant Cluster Head
Faculty: Theses from Anglia Ruskin University
Depositing User: Lisa Blanshard
Date Deposited: 29 Oct 2018 10:19
Last Modified: 29 Oct 2018 10:19
URI: http://arro.anglia.ac.uk/id/eprint/703767

Actions (login required)

Edit Item Edit Item