Hybrid Hierarchical Agglomerative Clustering Protocol for Wireless Sensor Networks

Clustering is one of the most energy-efficient ways to organize sensor nodes in Wireless Sensor Networks (WSNs). To perform clustering, location data are normally used for calculating the distance between sensor nodes. But location data may not always be available due to GPS failures or consideration of cost. Alternatively, Received Signal Strength (RSS) or RSS Indicator (RSSI) is used as the distance estimator, but it has been proved that RSS or RSSI is unreliable in many studies. In order to mitigate these problems, this thesis proposes a hybrid clustering protocol — Hybrid Distributed Hierarchical Agglomerative Clustering (H-DHAC) protocol which uses both quantitative location data and qualitative connectivity data in clustering for WSNs. Our simulation results show that H-DHAC has a lower percentage of compromise in performance in terms of network life time and total transmitted data compared to similar approach that uses complete location data. Further, it still outperforms the well known clustering protocols, e.g. LEACH and LEACH-C.