A Deep Learning Approach for Efficient Anomaly Detection in WSNs

Data reliability in Wireless Sensor Networks (WSNs) has a substantial influence on their smooth functioning and resource limitations. In WSN, the data aggregated from clustered sensor nodes are forwarded to base station for analysis. Anomaly Detection (AD) focuses detecting outlier ensure consistency during aggregation. As WSNs have critical limitations concerning energy consumption node lifetime, AD is supposed provide integrity with minimum consumption, which been an active research problem. Hence, researchers striving methods improve accuracy of handled concern constraints WSNs. This paper introduces Feed-forward Autoencoder Neural Network (FANN) model detect abnormal instances improved reduced consumption. The proposed also acts as False Positive Reducer intending reduce false alarms. It compared other dominant unsupervised algorithms over robustness significant metrics real-time datasets. Relatively, our yields fewer alarms thereby supporting sustainable WSN.