RF Jamming BERT Intrusion Detection Systems for Vehicular Networks
Bibliographic record
Abstract
As vehicular networks become increasingly complex and interconnected, they face rising security threats, particularly Radio-Frequency (RF) jamming attacks that can disrupt communication and compromise safety. While existing Artificial Intelligence (AI) models have been applied for RF jamming detection, they often face challenges in addressing the dynamic nature of vehicular environments. In this paper, we present an intrusion detection system (IDS) utilizing Bidirectional Encoder Representations from Transformers (BERT), a transformer-based Large Language Model (LLM) renowned for its contextual understanding. During the training of BERT, we incorporated and experimented with two tokenizing techniques: WordPiece and Byte-Pair Encoding (BPE). The results from our experiments highlighted that the BERT model fine-tuned with the WordPiece tokenization method significantly outperformed existing baseline methods. Specifically, this approach achieved a high accuracy of 96.25% at 25m/s and 90.25% at 15m/s showcasing strong performance gains compared to baseline methods: K-Nearest Neighbors (KNN), Random Forest (RaFo), Long Short-Term Memory (LSTM) that reached, respectively, 94.46% and 82.27%; 94.61% and 80.04%; 95.17% and 84.83% for 25 m/s and 15 m/s. The BPE tokenization IDS version achieved an accuracy of 85.00% at 25m/s and 80.25% at 15m/s. Our results demonstrate the effectiveness of the BERT RF Jamming IDS with WordPiece tokenization to detect and classify RF threats on vehicle networks.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".