Fault Prediction of IoT Terminals based on Improved ResNet and BiLSTM Models

Abstract

With the rapid development of the IoT business, the IoT is showing a trend of large-scale and complex, and the types and quantities of terminal devices connected to the IoT system are constantly increasing, which puts forward higher requirements for the stability of the IoT. At present, the fault of IoT terminal device is unavoidable, and the existing research in the field of IoT terminals fault mainly focuses on the monitoring and diagnosis of faults. It is particularly important to make accurate and timely prediction before the fault occurs. In this paper, a IoT terminal fault prediction algorithm based on improved ResNet and BiLSTM and a Knowledge Review algorithm based on ECA module and Channel Connection loss are are proposed, which provides an effective solution for fault prediction of IoT terminal device.