Decision Fusion for Clustered WSNs over IoT Infrastructure using 1-bit Quantization

This paper considers the decision fusion problem for wireless sensor networks (WSNs) where the decisions are transmitted to the fusion center (FC) over Internet-of-Things (IoT) infrastructure. The WSN is configured that each group of sensors transmit their decisions about a certain phenomenon to a cluster-head, and all cluster heads forward the decisions to the a remote FC that is hosted on the cloud. To improve the system spectral and power efficiency, we use 1-bit quantization at the sensor and cluster-head levels. Moreover, 1-bit censoring based quantizers are proposed where the cluster-head forwards only the reliable sensing data to the FC, while dropping the unreliable data. At the FC, the received data are fused to obtain the final decision. The obtained numerical results evaluated for several operating scenarios show that the 1-bit quantization process can produce reliable global decisions while saving the power and bandwidth.