Energy efficiency is an important issue during the design and the overall performance evaluation of an UWSN system. Clustering sensor nodes have proven to be an effective method to improve the load balancing and scalability of the network while minimizing the system's overall energy consumption. In this paper, a new clustering algorithm is proposed to provide an improved cluster system against cluster-head failures. This study suggests that system CH failures could be further minimized when simultaneously a CH (primary CH) and a vice/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 operates as a head node. This study proposes two major procedures in order this to be accomplished, the detection failure and the recovery procedures. The first one initially detects any failures that occurred in the network and then reports this information to the relevant nodes to initiate recovery. The recovery procedure actually decides who and when will trigger the recovery function according to the origin of the CH node failure which can be either the energy depletion of the CH's battery or a software/hardware malfunction. The simulation results clearly indicate that there is an improvement in terms of network lifetime expectancy and energy consumption.
展开▼
