Efficient in-network aggregation for wireless sensor networks with fine grain location-based cluster
Maznah Bte Kamat, Abdul Samad Ismail, Stephan Olariu · 2011
Ambient data generated by sensor nodes must be forwarded to sink node and made available to central system for further processing. Additionally, in a densely deployed sensor network, data are correlated and redundancy may occur. To minimize energy consumption, data may be aggregated during forwarding process. Although clustering helps in eliminating redundancy and prolonging network lifetime, but the formation and maintaining of cluster heads is energy consuming. Utilizing fine-grain location-based cluster (enhanced Virtual Infrastructure - eVI) [2] we demonstrate that clustering without cluster heads can provide energy efficient in-network aggregation and offers inherent default aggregators for fault-tolerance. Centralized routing for eVI cannot be done implicitly as in Virtual Infrastructure (VI) [1]. Each fine grain clusters must be scheduled properly during data forwarding in reducing energy usage due to data collision. Exploitation of mobile sink requires non-centralized routing in minimizing delay and data loss. Hybrid routing is proposed in which dynamic destination cluster is determined as a data collection point. Simulation results demonstrate that the in-network processing of clustering without cluster head performs better than cluster with a head.