ABSTRACT

With the emergence of cooperative communications for cellular networking, attention has been paid to clustering techniques, especially clustering algorithms. In this paper, we firstly simply introduce the state-of-the-art in Wireless Sensor Networks (WSNs) and clustering algorithms, and expound classical clustering protocols and the principle of the conventional K-means algorithm. Next, we propose a new clustering algorithm for WSNs based on the conventional K-means algorithm. Lastly, simulation studies show that our proposed clustering algorithm can adaptively get the number of clustering K, furthermore, the clustering result is more uniform, and the improved K-means algorithm is more suitable for high density network scenarios than the conventional one.