BiLCNet : BiLSTM-Conformer Network for Encrypted Traffic Classification with 5G SA Physical Channel Records

2025-09-22

AI Summary

To address the challenge of encrypted traffic classification in 5G standalone (SA) networks, this paper proposes an end-to-end classification method leveraging over-the-air physical-layer channel data. To overcome the limitations of conventional port-based and deep packet inspection (DPI) techniques in encrypted and dynamic scenarios, we design BiLCNet—a novel hybrid architecture integrating bidirectional long short-term memory (BiLSTM) and Conformer—marking the first application of Conformer to physical-layer traffic classification. BiLCNet jointly models temporal dependencies and local-global spatial features. Coupled with physical-channel-specific preprocessing and time-frequency domain feature engineering, the model achieves 93.9% accuracy on a real-world, noise-constrained 5G SA dataset—substantially outperforming baseline methods. Moreover, it demonstrates strong zero-shot transferability and cross-environment robustness.

Technology Category

• Machine Learning: Efficient ML / Green AI
• Natural Language Processing: Ethics — Bias, Fairness, Transparency & Privacy
• Computer Vision: Large Vision Models

Application Category

• Security and Privacy: Privacy-enhancing technologies
• Responsible Web: Measurement, analysis, and circumvention of Web censorship
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Accurate and efficient traffic classification is vital for wireless network management, especially under encrypted payloads and dynamic application behavior, where traditional methods such as port-based identification and deep packet inspection (DPI) are increasingly inadequate. This work explores the feasibility of using physical channel data collected from the air interface of 5G Standalone (SA) networks for traffic sensing. We develop a preprocessing pipeline to transform raw channel records into structured representations with customized feature engineering to enhance downstream classification performance. To jointly capture temporal dependencies and both local and global structural patterns inherent in physical channel records, we propose a novel hybrid architecture: BiLSTM-Conformer Network (BiLCNet), which integrates the sequential modeling capability of Bidirectional Long Short-Term Memory networks (BiLSTM) with the spatial feature extraction strength of Conformer blocks. Evaluated on a noise-limited 5G SA dataset, our model achieves a classification accuracy of 93.9%, outperforming a series of conventional machine learning and deep learning algorithms. Furthermore, we demonstrate its generalization ability under zero-shot transfer settings, validating its robustness across traffic categories and varying environmental conditions.

Problem

• Classifying encrypted traffic using 5G physical channel data
• Overcoming limitations of traditional port-based and DPI methods
• Capturing temporal dependencies and structural patterns in wireless data

Innovation

• BiLSTM-Conformer hybrid architecture for classification
• Preprocessing pipeline with customized feature engineering
• Uses 5G SA physical channel records as input