No abstract

No abstract

No abstract

Wireless sensor networks require efficient methods to identify anomalous sensor nodes and avoid being misled by the abnormal or wrong data. A novel approach is proposed to detect sensor node anomaly based on the statistical correlation in the sensing data from the sensor network. The proposed polynomialtime algorithm analyzes the correlations among different measurements in the history data of the normal sensor network. Correlation discriminant rules are learned and the anomaly discriminant rules are induced at the end of the algorithm. The anomaly discriminant rules are then used by any data fusion node in the network to find anomaly suspicious sensor nodes if the correlations in the recent sensing data satisfy any of the anomaly discriminant rules.

The optical spectra and EPR parameters of trigonal Yb3