Salvato in:
| Autore principale: | |
|---|---|
| Natura: | Recurso digital |
| Lingua: | inglese |
| Pubblicazione: |
Zenodo
2026
|
| Soggetti: | |
| Accesso online: | https://doi.org/10.5281/zenodo.19659074 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866902251225743360 |
|---|---|
| author | Eunhan |
| author_facet | Eunhan |
| contents | <p>Artifact for the paper "MIRAGE: Detecting Fake Emergency Electronic Brake Light Attacks in V2X Networks via Event-Gated Behavioral Analysis" (VehicleSec 2026).</p> <p><br>WHAT IS MIRAGE?</p> <p>MIRAGE is an event-gated misbehavior detector for fake Emergency Electronic Brake Light (EEBL) attacks in V2X networks. It activates only when an EEBL is received and evaluates four stages of evidence: context assessment, deceleration plausibility, behavioral consistency, and post-stop consistency. An additional IDM-based mitigation layer bounds reactive braking as a fail-safe for any attacks that evade detection.</p> <p><br>KEY RESULTS</p> <p>Aggregate results at threshold Theta=0.55 (5 seeds):</p> <p>- B1 Threshold: TPR=0.597, FPR=0.009, Precision=0.950, F1=0.733<br>- B2 VCADS: TPR=0.499, FPR=0.040, Precision=0.778, F1=0.607<br>- B3 F2MD/VASP: TPR=0.110, FPR=0.009, Precision=0.770, F1=0.193<br>- MIRAGE: TPR=0.883, FPR=0.009, Precision=0.965, F1=0.922</p> <p>MIRAGE is the only detector with high detection on both attack variants: 77.0% on A1 (NoStop) and 99.7% on A2 (WithStop). The IDM mitigation layer reduces mean braking severity by approximately 75%.</p> <p><br>CLEAN-ROOM VERIFIED REPRODUCIBILITY</p> <p>Full pipeline verified in a fresh ubuntu:20.04 Docker container (clone, install, run, verify):</p> <p>- Single-seed verify: 4/4 claims PASS, F1=0.9225<br>- Multi-seed verify (5 seeds): 4/4 claims PASS, F1=0.9223<br>- Paper reference: F1=0.9220<br>- Delta: 0.0003 (well within tolerance)</p> <p>All 10 paper figures are reproduced. 90/90 multi-seed configurations complete successfully. Full pipeline wall clock: approximately 1 hour 55 minutes.</p> <p><br>BADGES CLAIMED</p> <p>- Artifacts Available<br>- Artifacts Functional<br>- Results Reproduced</p> <p><br>BUILT ON</p> <p>- OMNeT++ 5.6.2<br>- Veins 5.2<br>- SUMO 1.8.0<br>- VASP (V2X Application Spoofing Platform)</p> <p><br>QUICK START</p> <p>Starting from a fresh Ubuntu 20.04 host:</p> <p> git clone --branch artifact-evaluation https://github.com/eunhanka/mirage-eebl-detector.git<br> cd mirage-eebl-detector<br> sudo ./scripts/install.sh all<br> ./scripts/quick_test.sh<br> ./scripts/run_single_seed.sh 0<br> ./scripts/verify_results.py --seeds 0</p> <p>For full 5-seed reproduction (approximately 80 minutes additional):</p> <p> ./src/vasp/scenario/run_multi_seed.sh 0 1 2 3 4<br> ./scripts/verify_results.py --seeds 0 1 2 3 4</p> <p>See README.md and CLAIMS.md in the repository for details and tolerances.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19659074 |
| institution | Zenodo |
| language | eng |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | eunhanka/mirage-eebl-detector: VehicleSec 2026 AE Final Submission (v2) Eunhan V2X misbehavior detection EEBL vehicular network security connected vehicles VehicleSec intelligent driver model VASP <p>Artifact for the paper "MIRAGE: Detecting Fake Emergency Electronic Brake Light Attacks in V2X Networks via Event-Gated Behavioral Analysis" (VehicleSec 2026).</p> <p><br>WHAT IS MIRAGE?</p> <p>MIRAGE is an event-gated misbehavior detector for fake Emergency Electronic Brake Light (EEBL) attacks in V2X networks. It activates only when an EEBL is received and evaluates four stages of evidence: context assessment, deceleration plausibility, behavioral consistency, and post-stop consistency. An additional IDM-based mitigation layer bounds reactive braking as a fail-safe for any attacks that evade detection.</p> <p><br>KEY RESULTS</p> <p>Aggregate results at threshold Theta=0.55 (5 seeds):</p> <p>- B1 Threshold: TPR=0.597, FPR=0.009, Precision=0.950, F1=0.733<br>- B2 VCADS: TPR=0.499, FPR=0.040, Precision=0.778, F1=0.607<br>- B3 F2MD/VASP: TPR=0.110, FPR=0.009, Precision=0.770, F1=0.193<br>- MIRAGE: TPR=0.883, FPR=0.009, Precision=0.965, F1=0.922</p> <p>MIRAGE is the only detector with high detection on both attack variants: 77.0% on A1 (NoStop) and 99.7% on A2 (WithStop). The IDM mitigation layer reduces mean braking severity by approximately 75%.</p> <p><br>CLEAN-ROOM VERIFIED REPRODUCIBILITY</p> <p>Full pipeline verified in a fresh ubuntu:20.04 Docker container (clone, install, run, verify):</p> <p>- Single-seed verify: 4/4 claims PASS, F1=0.9225<br>- Multi-seed verify (5 seeds): 4/4 claims PASS, F1=0.9223<br>- Paper reference: F1=0.9220<br>- Delta: 0.0003 (well within tolerance)</p> <p>All 10 paper figures are reproduced. 90/90 multi-seed configurations complete successfully. Full pipeline wall clock: approximately 1 hour 55 minutes.</p> <p><br>BADGES CLAIMED</p> <p>- Artifacts Available<br>- Artifacts Functional<br>- Results Reproduced</p> <p><br>BUILT ON</p> <p>- OMNeT++ 5.6.2<br>- Veins 5.2<br>- SUMO 1.8.0<br>- VASP (V2X Application Spoofing Platform)</p> <p><br>QUICK START</p> <p>Starting from a fresh Ubuntu 20.04 host:</p> <p> git clone --branch artifact-evaluation https://github.com/eunhanka/mirage-eebl-detector.git<br> cd mirage-eebl-detector<br> sudo ./scripts/install.sh all<br> ./scripts/quick_test.sh<br> ./scripts/run_single_seed.sh 0<br> ./scripts/verify_results.py --seeds 0</p> <p>For full 5-seed reproduction (approximately 80 minutes additional):</p> <p> ./src/vasp/scenario/run_multi_seed.sh 0 1 2 3 4<br> ./scripts/verify_results.py --seeds 0 1 2 3 4</p> <p>See README.md and CLAIMS.md in the repository for details and tolerances.</p> |
| title | eunhanka/mirage-eebl-detector: VehicleSec 2026 AE Final Submission (v2) |
| topic | V2X misbehavior detection EEBL vehicular network security connected vehicles VehicleSec intelligent driver model VASP |
| url | https://doi.org/10.5281/zenodo.19659074 |