Research On Lightweight Traffic Anomaly Detection Technology In Ubiquitous Electric Internet Of Things

The ubiquitous power Internet of Things is a professional expression of the wide application of Internet of Things technology in the power industry.With the in-depth development of Internet+ ubiquitous power Internet of Things research and construction,the threat of attacks on power networks is becoming more and more serious,which has brought great pressure to the security of ubiquitous power Internet of Things.Network traffic anomaly detection can detect abnormal traffic and potential attacks in time through the processing and analysis of network traffic data.With the maturity and wide application of deep learning technology,certain results have been obtained in the field of network traffic anomaly detection.Based on the ubiquitous power Internet of Things terminal equipment,this paper studies traffic anomaly detection technology based on lightweight deep learning algorithms.The research work of this paper mainly includes the following aspects:(1)On the basis of fully understanding the development status of Ubiquitous Power IoT and terminal security requirements,a lightweight flow anomaly detection task suitable for Ubiquitous Power IoT terminals is proposed;(2)By analyzing the respective characteristics of the currently commonly used flow anomaly detection technologies based on deep leaning,choose a lightweight convolutional neural network method that can not only reduce the computational burden but also ensure the detection effect,and the MobileNet V2 model was selected for research;(3)Aiming at the one-dimensional time series characteristics of flow data,an improvement plan is proposed.On the basis of the commonly used two-dimensional model architecture,the dimensions of network operations involved were reduced.Use the MobileNet V2 model to experiment.First,preprocess the KDD99 data set selected in the experiment,divide it into training set and validation set according to the ratio of 8:2,Then train the training set and the verification set separately and verify the model effect through the results of the verification set,and finally evaluate the model performance based on the comprehensive evaluation indicators such as F1 scores.The experimental results show that the one-dimensional MobileNet V2 model can well complete the task of flow anomaly detection,and can well adapt to the needs of the Ubiquitous Power IoT.