Multi-Sensor Data Fusion for Cluster-based Data Aggregation in IoT Applications
Surender Redhu, Rajesh Mahanand Hegde · 2019
Clustering of devices plays a significant role in the development of energy-efficient wireless sensor networks for IoT applications. In this paper, a novel clustering method is proposed which utilizes the multiple sensor modalities of devices to develop an energy-efficient data aggregation protocol. It groups the devices of a network into different clusters using a fused resemblance matrix. The proposed method fuses different sensor modalities like radio, acoustic and light to obtain robust resemblance coefficients. The weights for fusion of different modalities are computed using the modelling error and within a confidence interval. The proposed clustering method is developed using the hierarchical agglomerative clustering framework. Performance of the proposed method in cluster-based data aggregation is evaluated for various sensor networks. Experimental results illustrate the significance of the proposed multi-sensor data based clustering scheme in energy-efficient data aggregation over IoT networks.