Table 1 Overview of existing work.
From: A hybrid deep learning model for detection and mitigation of DDoS attacks in VANETs
Author(s) | Year | Methodology | Dataset | Advantages | Disadvantages |
|---|---|---|---|---|---|
Bhanja et al.44 | 2020 | Fuzzy logic controllers for Sybil and DDoS detection | Not specified | Improved accuracy, sensitivity, and recall for attack detection | No details on scalability to larger VANET networks |
Dhar et al.45 | 2021 | CascadMLIDS using cascaded ML with PCA | NSL-KDD | Increased reliability and precision for intrusion detection | Complexity due to cascaded framework |
Verma et al 46 | 2019 | PREVIR with Logit and LogitBoost | KDDCup’99 | High accuracy (99.99%) and 100% true positive ratio | High average false positive ratio (35%) |
Amaouche et al.47 | 2022 | FSCB-IDS with feature selection and class imbalance handling | CIC-IDS2017 | Efficient feature selection, effective class imbalance handling | No information on real-time performance |
Alsarhan et al.48 | 2021 | SVM with GA, PSO, and ACO | UNSW-NB15 | Improved predictive capabilities, reduced dimensionality dependence | High complexity due to optimization techniques |
Rashid et al.49 | 2022 | Distributed multi-layer classifier with AWS integration | Custom VANET Dataset | Real-time classification, high accuracy (up to 99%) | Scalability issues with increasing nodes in the network |
Sontakke and Chopade50 | 2023 | Deep learning with autoencoder and Beetle-Whale Swarm Optimization | NSL-KDD | Enhanced security, effective trust-based routing | Computational cost of feature selection and training |
