Research On Security Detection Technology Based On LTE Air Interface Protocol

By 2020,LTE network coverage in my country has reached 98%.Today,when 4G is increasingly popular,network security is particularly important.An air interface is an air interface exposed to the outdoors,accessible to anyone.Therefore,the air interface is the weakest interface that is most easily controlled and attacked by attackers.This article focuses on attack detection on air-based denial of service attacks and replay attacks.In response to LTE air interface attackers sending a large number of illegal information,causing illegal users to enter the access authentication process,making the HSS(Home Subscriber Server)core network occupy a large amount of computing resources,resulting in the detection of denial of service attacks that normal users cannot access.This article analyzes the access authentication process from the perspective of signaling changes,and extracts 6 attribute features based on the data changes in this attack scenario.Use Bi?LSTM neural network to detect denial of service attacks.The experimental results of this paper not only compare the accuracy rate,false negative rate,and false report rate of LSTM network and Bi?LSTM network,but also compare the above three aspects from the perspectives of data preprocessing,feature selection,and the number of network layers.Finally,it is determined that the four-layer Bi?LSTM network after adding data preprocessing and feature selection has the highest accuracy rate of 92.9%.For LTE air interface attackers,intercept the terminal's last handover service signaling and resend it after a period of time to deceive the system and interfere with the terminal's replay attack detection.This article still analyzes the switching process that is most prone to this attack from the perspective of signaling changes,and extracts relevant attribute characteristics based on the signaling changes of the switching process in the attack scenario,abandoning the traditional methods of adding timestamps and random numbers,and is innovative A classifier based on the improved Naive Bayes algorithm is built to detect replay attacks.The algorithm not only comprehensively considers the independence and dependence between feature attributes through structural improvement,but also uses genetic algorithm to solve the optimal combination of attributes for category parameters.At the end of the article,the experimental results are compared based on the algorithm before and after the improvement,and the accuracy of the improved algorithm is as high as 95.13%.