Improvements in LEACH based on K-means and Gauss algorithms

Several methods have been developed for routing in wireless sensor networks; the most popular is LEACH protocol. It owes its popularity to its simplicity and ability to handle large data sets. However, the main limitation of this method is its dependency to probabilistic calculations. In this paper, we propose a new improved version of the LEACH protocol. Our main contribution consists on the integration of an improved unsupervised algorithm which is the K-means on clusters and the use of the Gaussian elimination algorithm for choosing the Cluster Head. Based on Davies–Bouldin index (DBI), the K-means finds the optimal number of clusters K. The Davies-Bouldin index makes calculations much less complex, which is a potential purpose in clustering. Thus, DBI based K-means offers more simple parallel distributed computing of K and allows faster auto-classification of sensor nodes. In this manner a balanced energy clusters is created and distributed energy consumption is guaranteed. Simulations in a wireless sensor network show that our new approach greatly reduces the computation time thanks to K-means and minimize energy consumption thanks to Gauss.