Energy efficient cluster based algorithm for underwater sensor networks

  • Kyriakos Ovaliadis

    Student thesis: Doctoral Thesis


    In this thesis, an innovative and evolving Cluster Based Routing Algorithm (CBRA) is proposed to provide an improved energy efficiency cluster system which can also capable of handling cluster-head and mobile sensor node connectivity failures.

    In addition, to develop, implement and test CBRA, a new simulator called USNeT (Underwater Sensor Network simulation Tool) has been designed, developed and implemented. This USNeT simulator follows the object-oriented design style and all network entities are implemented as classes in the C++, encapsulating thread mechanisms. Initially significant adjustments have been made in order for the algorithm to become more energy efficient. Some of these alterations are: transmission range management, re-cluster process activation for each group separately, sensor node sleeping mode and unwanted information rejection. All the simulation results which were implemented against Low Energy Adaptive Clustering Hierarchy (LEACH) protocol indicate a small but significant improvement in the performance of the CBRA especially in energy efficiency.
    This study also suggests that system Cluster Head (CH) failures could be further minimized when simultaneously a CH (primary CH) and a backup CH are selected. Thus, when a primary CH fails due to an irreparable fault, a backup CH will take its place and it will operate as a head node. Therefore, the CBRA is redefined and optimised to be able to handle this issue and also to diminish any communication link establishment interruptions. The analysis of the simulation results shows that the redefined CBRA (r-CBRA) is more energy efficient and can effectively enhance the network survivability capacity in the event of cluster-head failures, than the scheme with the non-optimised algorithm CBRA and the LEACH protocol.

    Thereafter, the r-CBRA is used again, to address sensor node connectivity failures. In case of a mobile sensor node that is close to a cluster but not in the range of a CH, r-CBRA changes the status of the nearest sensor node to a CH and then it establishes a communication link between them. Simulation results show once more that the new cluster based routing algorithm ensures the connectivity of the network without sacrificing the energy efficiency of the network.

    Date of AwardFeb 2015
    Original languageEnglish
    SupervisorNick Savage (Supervisor), David Ndzi (Supervisor) & Spencer Onuh (Supervisor)

    Cite this