Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 42 of the 42 references it cites were found in a registry or an index; 2 of the 10 sources its claims cite were checked for support.

X-GEVON- a novel explainable intelligent network to detect the multiple attacks in vanet systems

Open access
Anjali, Thuvva, Goyal, Rajeev, Balaji, G. N. · 2025 · Discover Computing

Abstract

Vehicular Ad-Hoc Networks (VANETs) play a vital role in the advancement of Intelligent Transportation Systems (ITS), offering secure and efficient communication between vehicles to support smart and traffic-independent road systems. With the integration of Internet of Things (IoT) technologies, vehicles have become capable of exchanging critical information over the internet, enabling real-time decision-making. However, this rapid digitization has also introduced significant security and privacy vulnerabilities, leading to increased susceptibility to malicious attacks that may cause catastrophic consequences. To address these security concerns, this study aims to develop a real-time, lightweight, and efficient deep learning model for detecting intrusions in VANET environments. The proposed framework, named X-GEVON, combines Gated Recurrent Units (GRU) for temporal feature extraction with an Enhanced Energy Valley Optimization (EVO) algorithm for hyperparameter tuning, ensuring high detection performance with minimal computational overhead. Furthermore, Explainable Artificial Intelligence (XAI) techniques, particularly the Local Interpretable Model-Agnostic Explanations (LIME), are integrated to provide transparent and interpretable classification results. The model is trained and tested on approximately 400,000 real-time data traces, including both normal and attack scenarios, generated using simulation tools such as SUMO, OMNET + + , and Python 3.19. Experimental results demonstrate that the proposed method significantly outperforms existing deep learning techniques, achieving a detection accuracy of 99.6%, a precision of 99.2%, and a recall of 99.4%. Additionally, the use of LIME enhances the interpretability of the model by explaining its prediction logic, making it more reliable for real-world applications. In conclusion, the X-GEVON framework offers a powerful, accurate, and explainable solution for intrusion detection in VANETs, bridging the gap between high-performance security models and transparent, interpretable AI systems.

Overview

CoChat Summary

TL;DR

The paper proposes X-GEVON, an explainable intrusion-detection framework for VANETs that combines modified GRU feature extraction, Enhanced Energy Valley Optimization for hyperparameter tuning, and LIME for interpretability. It reports high detection performance on approximately 400,000 simulated real-time traces, with abstract-level accuracy of 99.6%, precision of 99.2%, and recall of 99.4%.

Key findings

  • On binary normal/attack detection, X-GEVON reached 99.5% accuracy, 99.4% precision, 99.2% recall, 99.0% specificity, and 99.1% F1 at k=5; performance remained stable across k=1–5, with accuracy rising from 99.2 at k=1 to 99.5 at k=5.
  • Ablation showed adding EEVO to GRU improved accuracy from 91.2 to 99.5, precision from 90.98 to 99.4, recall from 90.8 to 99.2, specificity from 91 to 99.0, and F1-score from 90.0 to 99.1.
  • Multi-class results reported X-GEVON accuracy of 99.6% for DoS, 99.82% for MIM, and 99.7% for Sybil/probing attacks; the abstract reports overall 99.6% accuracy, 99.2% precision, and 99.4% recall.
  • Computational cost was lower than baselines: X-GEVON training time 155 s, inference time 1.4 ms/packet, and model size 5.2 MB, versus LSTM 204 s/3.5 ms/9.6 MB, GRU 190 s/3.0 ms/8.1 MB, and HHO-GRU 230 s/2.6 ms/8.9 MB. LIME indicated data rate had more impact on attack detection.

Methods

  • Design: X-GEVON has four components—Real-Time Data Generation, Data-Preprocessing Unit, GRU Feature Extraction, and Enhanced Energy Value Optimized Deep Learning Network; the modified GRU reduces trainable parameters by a factor of 2 × nm, EEVO tunes hyperparameters, and LIME provides interpretability.
  • Data: approximately 400,000 traces were generated using SUMO, OMNET++, VEINS, INET, and an Attack Injection Module; the route from Madurai to Hyderabad was simulated for 2 months; labels include normal, DoS, Blackhole, DDoS, and Man-in-the-Middle, while multi-class analysis also includes Sybil and probing attacks.
  • Analysis: dataset split was 70% training, 20% testing, and 10% validation; results were averaged over 180 epochs; k=5 validation was used; metrics included accuracy, sensitivity/recall, specificity, precision, and F1-score; statistical validation used fitness-function assessment over 50 iterations and Wilcoxon signed-rank testing.

Limitations

  • The paper does not state a limitations section for X-GEVON.
  • Evident from the text: evaluation relies on simulated traces generated by SUMO/OMNET++/VEINS with injected attacks, not real-world VANET deployment, and the data come from a single simulated route (Madurai to Hyderabad) over 2 months.
  • Evident from the text: some reported results are inconsistent or truncated—for example, the abstract reports 99.6% accuracy while the binary k=5 table reports 99.5%, and the DDoS X-GEVON row in Table 11 is incomplete—which limits reproducibility and interpretation.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s10791-025-09696-x

Cites

42 of 42 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 Model System of VANET
Fig. 1 Model System of VANETCC-BY-NC-ND · From the PDF
Fig. 2 X-GEVON- Its Complete Frameworks and Its working component
Fig. 2 X-GEVON- Its Complete Frameworks and Its working componentCC-BY-NC-ND · From the PDF
Fig. 3 Data Generation in Real Time Scenario using SUMO Environment (Linked with Satellite Environment)
Fig. 3 Data Generation in Real Time Scenario using SUMO Environment (Linked with Satellite Environment)CC-BY-NC-ND · From the PDF
Image 4
CC-BY-NC-ND · From the PDF
Fig. 5 Conversion of real time vehicles as the communication nodes in OMNET + + using VEINS interfaces
Fig. 5 Conversion of real time vehicles as the communication nodes in OMNET + + using VEINS interfacesCC-BY-NC-ND · From the PDF
Fig. 6 Injection of Attacks in OMNET + + using Attack injection Module
Fig. 6 Injection of Attacks in OMNET + + using Attack injection ModuleCC-BY-NC-ND · From the PDF
Fig. 7 Performance of the X-GEVON framework in detecting the normal instances at different drop-outs
Fig. 7 Performance of the X-GEVON framework in detecting the normal instances at different drop-outsCC-BY-NC-ND · From the PDF
Fig. 8 Performance of the X-GEVON approach in recognising the attack instances at different drop-outs
Fig. 8 Performance of the X-GEVON approach in recognising the attack instances at different drop-outsCC-BY-NC-ND · From the PDF
Image 9
CC-BY-NC-ND · From the PDF
Image 10
CC-BY-NC-ND · From the PDF
Fig. 10 Loss-validation curves a validation for the X-GEVON framework in recognising the attack instances. b Normal Scenario
Fig. 10 Loss-validation curves a validation for the X-GEVON framework in recognising the attack instances. b Normal ScenarioCC-BY-NC-ND · From the PDF
Fig. 11 Convergence analysis of the varied optimization approaches in detecting the varied attack instances
Fig. 11 Convergence analysis of the varied optimization approaches in detecting the varied attack instancesCC-BY-NC-ND · From the PDF
Fig. 12 Cumulative performance evaluation of the recommended X-GEVON at varied drop-outs to detect DoS, MIM and Sybil Attacks
Fig. 12 Cumulative performance evaluation of the recommended X-GEVON at varied drop-outs to detect DoS, MIM and Sybil AttacksCC-BY-NC-ND · From the PDF
Fig. 13 Cumulative Performance Analysis of the Proposed X-GEVON at different drop-outs to detect DDoS and Probing Attacks
Fig. 13 Cumulative Performance Analysis of the Proposed X-GEVON at different drop-outs to detect DDoS and Probing AttacksCC-BY-NC-ND · From the PDF
Fig. 14 Convergence analysis of different models in detecting the multiple attacks in VANET
Fig. 14 Convergence analysis of different models in detecting the multiple attacks in VANETCC-BY-NC-ND · From the PDF
Fig. 15 Spider plot diagram for the importance of features in the normal instance using LIMA model
Fig. 15 Spider plot diagram for the importance of features in the normal instance using LIMA modelCC-BY-NC-ND · From the PDF
Fig. 16 Spider diagram for the feature distribution in the single attack scenario (DoS)
Fig. 16 Spider diagram for the feature distribution in the single attack scenario (DoS)CC-BY-NC-ND · From the PDF
Fig. 17 Spider diagram for the feature distribution in the multiple attack scenario
Fig. 17 Spider diagram for the feature distribution in the multiple attack scenarioCC-BY-NC-ND · From the PDF

Verification

Claims the paper makes

2 Valid · 2 Not valid · 5 Not checked · 33 Own results and interpretations

18 more claims

Create a free account to see all 42 claims and to verify the 1 reference still open.

See all claims — free account
  • It may state two separate claims as one. 12
  • Probably not a substantial claim. 8
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

2 of 2

Cited works that were retracted

Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.

2 of 2

Citations

Cites

2 retracted

Show all 40 references

Research works printed but not linked: 2

  1. Miller C, Valasek C. Adventures in automotive networks and control units. Def Con. 2013;21(15–31):13.
  2. Maui, HI, USA, 4–7 November 2018, pp.2575–2582. New York: IEEE. 52. INET. INET framework, https://inet.omnetpp.org/. Accessed 14 July 2019.

Cited by

See all 6

Metadata

Authors
Anjali, T., Goyal, R., & Balaji, G. N.
Affiliations
Amity University Madhya Pradesh (IN)
Corresponding author
Anjali, Thuvva
Venue
Discover Computing
Published
2025-08-27
Volume
28
Issue
1
ISSN
2948-2992 (ISSN-L)
OpenAlex
W4414837887
Cited by
6
References
40
Citations by year
2026: 4 · 2025: 2
FWCI
1.35
Type
journal article
Language
en
Open access
Licence
Locations
  • Discover Computing (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Physical Sciences › Engineering › Electrical and Electronic Engineering › Vehicular Ad Hoc Networks (VANETs)
  • Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
  • Physical Sciences › Computer Science › Artificial Intelligence › Adversarial Robustness in Machine Learning
Keywords
vehicular ad hoc networks, intrusion detection, malicious attack detection, explainable AI, gated recurrent unit, hyperparameter optimization, lime, temporal feature extraction, lightweight deep learning, network security, intelligent transportation systems

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: X-GEVON- a novel explainable intelligent network to detect the multiple attacks in vanet systems. https://app.cochat.ai/papers/10.1007/s10791-025-09696-x
The paper
Anjali, T., Goyal, R., & Balaji, G. N. (2025). X-GEVON- a novel explainable intelligent network to detect the multiple attacks in vanet systems. Discover Computing. https://doi.org/10.1007/s10791-025-09696-x

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.