A Spatiotemporal and Multivariate Attribute Correlation Extraction Scheme for Detecting Abnormal Nodes in WSNs

Many heterogeneous sensors exhibit a strong correlation in both space and time that can be used to enhance the abnormal node detection problem wireless sensor network (WSN). Corruption these correlations has been shown effective detecting false data injection attacks. However, extracting such is considered challenging task WSNs. In this paper, we adopt new cross-correlation-based method extract relationships. It utilizes observed spatiotemporal (ST) multivariate-attribute (MVA) decide whether subject abnormalities or represents actual events. Based on ST correlations, cross-correlation extracted by conducting shape-based logical subclustering two-phase analysis methods. first phase, system uses variable-size sliding window median absolute deviation (MAD) measure. If collected streams output certain percentage of anomalous points, MAD will flag points as measurements, all size passed second phase. The latter performs tumbling-window sliding-window analyses multicriteria measures. Finally, time-series features fed clustering generate similarity-like graph. reflects similarity degree with other nodes. nodes low below threshold identified candidate MVA set consisting few rules introduced check detected represent if exist, then are reported. Our experiments using two real-world datasets demonstrate our proposed approach detects an average accuracy 96.50%, precision 88.69%, recall rate 93.00%.