Abstract:

Wireless sensor networks (WSN) have become a real-time environmental monitoring solution and are widely used in various fields.Since the sensors in the network are easily affected by complex working environment,their own hardware and other factors,they may fail to work.Therefore,fault detection in wireless sensor networks is an indispensable link in its application field.To address the problem of fault sensor detection in the wireless sensor network,this paper proposes a fault detection model named GCN-GRU,which hybridizes a graph convolutional network (GCN) and a gate recurrent unit (GRU).The model consists of three layers:input layer,spatiotemporal processing layer and output layer.The input layer receives the sensor network data and the graph model constructed by the WSN and transmits them to the spatiotemporal processing layer.In the spatiotemporal processing layer,the spatial distribution features of the WSN and the characteristics of faults in high-dimensional space are extracted by the GCN,and they are constructed as the high-dimensional data of time series which act as the input of the GRU.Then the temporal evolution features of sensor network data and the temporal and spatial evolution characteristics are extracted and fused by the GRU.Finally,the fault detection results are obtained in the output layer.To evaluate the performance of the GCN-GRU,this paper compares the GCN-GRU model with existing fault detection algorithms for the WSN.Numerical experiments show that the GCN-GRU model can significantly improve the fault detection rate and reduce the false alarm rate,thus effectively identifying faulty sensors compared with the existing algorithms.

Key words: wireless sensor network, fault detection, graph convolutional network, gate recurrent unit