Optimized Self Organized Sensor Networks

Abstract

Wireless sensor networks are composed of a huge number of sensor nodes, which have limited resources - energy, memory and computation power. Energies are directly related to the lifetime of sensor network. If sensor nodes can be grouped to clusters, cluster member sensor nodes only need to communicate with cluster center (head) and this leads to energy conservation of the member sensors. So, how to compose clusters with minimal number of cluster heads, while including each node in a cluster is an important research issue. We propose a new advanced optimization algorithm for sensor network clustering. Using the proposed optimization algorithm, redundant cluster heads are eliminated, and unnecessarily overlapped clusters are merged. Optimization algorithm can be used as a clustering algorithm by itself and also manage the dynamic changes like node addition or die-out, while the network is even on the working state. We tested the proposed method as a clustering algorithm and compared it with two other recent sensor network clustering algorithms, Algorithm for Cluster Establishment (ACE) and Self Organizing Sensor network algorithm (SOS). The experiments results not only illustrate that the proposed algorithm could result in clusters with smaller number of cluster heads than others with any density of sensor networks, but also that the performance is more stable, which is also verified through repeated experiments.

Keywords

• Optimization of clustering algorithm
• Self Organizing Sensor algorithm
• Intelligent Clustering
• Wireless Sensor Network