An Energy Aware Adaptive Kernel Density Estimation Approach to Unequal Clustering in Wireless Sensor Networks.

Fagui Liu,Yufei Chang
DOI: https://doi.org/10.1109/ACCESS.2019.2902243
IF: 3.9
2019-01-01
IEEE Access
Abstract:Energy conservation is one of the most important challenges in wireless sensor networks (WSNs). Therefore, compared with the traditional networks, the WSNs not only need high-quality services with high throughput or low transmission delay, but also pay greater attention to energy utilization to extend network lifetime. The clustering routing algorithm is considered to be among the effective ways to collect and transmit data in WSNs. Cluster head (CH) plays a vital role in the cluster which is in charge of data aggregation and data transmission, so their energy consumption is higher than non-CH nodes. The traditional clustering algorithm tends to have the same size in each cluster. However, due to the randomness of the node distribution, the equal clustering mechanism obviously cannot reduce energy consumption. In order to solve this problem, this paper contributes a new unequal clustering algorithm, an energy-aware adaptive kernel density estimation algorithm (EAKDE), which aims to balance the energy dissipation among the CHs. EAKDE utilizes fuzzy logic to determine the priority of nodes competing for CH. In order to adapt the dynamic change of node conditions, adaptive kernel density estimation algorithm is utilized to assign the appropriate unequal cluster radius to sensor nodes. The simulation results demonstrate that, in different scenarios, EAKDE outperforms the other well-known algorithms in terms of network stability, network lifetime, and energy efficiency.
What problem does this paper attempt to address?