.png){kind=icon}
.png){kind=icon}
Return
5G Native Network Function for False Base Station Detection Using Machine Learning Technique
.png){kind=icon}
PRE
AI
DOI：10.1007/s10796-025-10614-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The Fifth-Generation (5G) technology is increasingly recognized as a key enabler for V2X applications, although large-scale commercial deployments remain in the early stages. Despite its potential to support advanced vehicular connectivity, 5G systems are still vulnerable to false base station attacks. While various strategies have been proposed for detecting these attacks, they cannot easily be adapted to the 5G ecosystem due to their limited applicability to specific vendors and limited operational knowledge of the system. To address this issue, we propose a new network function called the Measurement Report (MR) Verification Function (MRVF) which can detect false base stations and is compatible with the 5G network infrastructure. MRVF gathers MR messages from User Equipment and learns a model for detecting false base stations through Machine Learning. We simulated two scenarios: one with a normal environment, and another that included the presence of a false base station. We used class-labeled data gathered from each simulation for supervised learning. The generated classification model detected false base stations with different scene behavior in the subsequent simulation experiment. Among the seven supervised learning algorithms evaluated in this study, the most suitable algorithm according to the results was the Multi Layer Perceptron, which achieved a detection accuracy of 99% when the noise level was 0%.
Keywords:
False base station
5G Security
Machine learning
Supervised learning
Vehicle-to-everything (V2X)
Journal
IF:
8.3
Papers:
2.0K
Citations:
6.5K
Organization
View more organizations