Data aggregation scheme using neural networks in wireless sensor networks

Energy efficiency is the most important issue in all facets of wireless sensor networks (WSNs) operations because of the limited and non-replenishable energy supply. And WSNs are deployed in environments where sensors can be exposed to conditions that might interfere with the sensor readings. Moreover, a variety of sensors may be attached to WSNs to monitor the environment. Data aggregation, eliminating the data redundancy and improving the accuracy of information-gathering, is essential for WSNs. Hence, BPNDA was proposed, a data aggregation scheme based on back-propagation network (BPN). In the BPNDA, a three-layer BP neural network was used. The input layer neurons are located in cluster members (CMs), while the hidden layer neurons and the output layer neurons are located in cluster head (CH). Only the extracted data that represented the features of the raw data will be transmitted to the sink, so the efficiency of data gathering is improved and the total energy consumption is reduced.