Performance of a clustering algorithm for high density wireless sensor networks

The present study investigates the performance of a clustering algorithm for sensor networks called SWEET (Slotted Waiting period Energy-Efficient Time-driven clustering algorithm) [1] under a variety of situations. The effectiveness of SWEET has been illustrated by changing both the number and the radio range of nodes considering several performance metrics such as number of selected Cluster Heads (CH) and the time delays of CH selections. It is shown that SWEET can generate desirable number of CHs within an acceptable time delay in densely deployed scenarios. Based on an energy saving model, it is observed that inter-cluster multihop routing schemes are more energy efficient compared to intra-cluster multihop routing schemes for relaying data in a network which is decomposed by SWEET. Further, it has been shown that provided the network density satisfies an analytical lower bound, the distance between adjacent CHs selected by SWEET is bounded, which alternately facilitate the inter-cluster multihop routing. Simulations on high density sensor networks show that SWEET performs satisfactorily in organizing nodes with varying radio ranges into various clusters with neighboring CHs being bounded by a particular distance.