Unsupervised K-means for Energy Conservation in IoT Networks
Malha Merah, Zibouda Aliouat · 2022
The K-means clustering algorithm is one potential solution that prolongs the lifetime of non-uniformly distributed wireless networks. However, this algorithm is sensitive to initialization, which, if unsuitable, can lead to inadequate clustering. This paper proposes the implementation of Unsupervised K-means (U-K-means), a free initialization version of K-means, as a new clustering technique for topology management in wireless sensor networks. The simulation results demonstrate that U-K-means outperforms the traditional K-means algorithm-based solutions in terms of consumed energy and network durability.