Clustering algorithms for sensor networks and mobile ad hoc networks to improve energy efficiency

Doctoral Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title

University of Cape Town

Many clustering algorithms have been proposed to improve energy efficiency of ad hoc networks as this is one primary challenge in ad hoc networks. The design of these clustering algorithms in sensor networks is different from that in mobile ad hoc networks in accordance with their specific characteristics and application purposes. A typical sensor network, which consists of stationary sensor nodes, usually has a data sink because of the limitation on processing capability of sensor nodes. The data traffic of the entire network is directional towards the sink. This directional traffic burdens the nodes/clusters differently according to their distance to the sink. Most clustering algorithms assign a similar number of nodes to each cluster to balance the burden of the clusters without considering the directional data traffic. They thus fail to maximize network lifetime. This dissertation proposes two clustering algorithms. These consider the directional data traffic in order to improve energy efficiency of homogeneous sensor networks with identical sensor nodes and uniform node distribution. One algorithm is for sensor networks with low to medium node density. The other is for sensor networks with high node density. Both algorithms organize the clusters in such a way that the cluster load is proportional to the cluster energy stored, thereby equalizing cluster lifetimes and preventing premature node/cluster death. Furthermore, in a homogeneous sensor network with low to medium node density, the clusterhead is maintained in the central area of the cluster through re-clustering without ripple effect to save more energy. The simulation results show that the proposed algorithms improve both the lifetime of the networks and performance of data being delivered to the sink. A typical mobile ad hoc network, which usually consists of moveable nodes, does not have a data sink. Existing energy-efficient clustering algorithms maintain clusters by periodically broadcasting control messages. In a typical mobile ad hoc network, a greater speed of node usually needs more frequent broadcasting. To efficiently maintain the clusters, the frequency of this periodic broadcasting needs to meet the requirement of the potentially maximum speed of node. When the node speed is low, the unnecessary broadcasting may waste significant energy. Furthermore, some clustering algorithms limit the maximum cluster size to moderate the difference in cluster sizes. Unfortunately, the cluster sizes in these algorithms still experience significant difference. The larger clusters will have higher burdens. Some clustering algorithms restrict the cluster sizes between the maximum and minimum limits. The energy required to maintain these clusters within the maximum and minimum sizes is quite extensive, especially when the nodes are moving quickly. Thus, energy efficiency is not optimized.

Includes bibliographical references (leaves 161-172).