一种基于时空相关性的传感器网络节点故障检测方法

Located on the under layer of Internet of Things (IoT), Wireless Sensor Networks (WSN) is the essential element for the successful operation of IoT. Since the hostile environment in which WSN nodes are implemented, it is high possibility that the WSN nodes would be damaged or destroyed. The data accuracy of WSN would be decreased due to the broken nodes. Through analyzing the feather of collected data, we proposed a faulty node detection method based on special-temporal correlation in this paper. The traditional detection method based on temporal correlation is simple but not accurate enough, while the detection method based on spatial correlation holds high accuracy and complexity. Combining the features of two detection methods, the method based on spatial-temporal correlation guarantees both simplicity and accuracy of detection. Meanwhile, we erase the inter-nodes communications in temporal correlation detection to save the communication cost, and involve the clustering algorithms to cluster the spatial nodes and increase their spatial correlation.