Large-scale heterogeneous terminal multi-domain joint fault diagnosis direction

The safe and stable operation of the Internet of Things (IOT) system is strongly related to the economy and daily life. IOT equipment damage accidents have caused major economic losses, which impacted on people's lives, and endanger industrial and social safety. However, accidents caused by various accidental factors cannot be completely avoided, effective monitoring and fault diagnosis of the IOT system, in particular in power system are very important. From the perspective of large-scale heterogeneous terminals and heterogeneous terminals across domains, this paper conducts fault diagnosis on the operation of the IOT based on big data and artificial intelligence methods. Firstly, based on the data-driven method, the information collected by different sensors is processed through the deep neural network for dimensionality reduction and feature extraction. Secondly, the extracted multi-source information features are feature fused. Finally, the Softmax function is used to diagnose the faults of the fused features.