Optimized stacked ensemble approach for detecting position falsification in VANETs

Area of Science:

• Computer Science
• Network Security
• Artificial Intelligence

Background:

• Vehicle Ad hoc Networks (VANETs) are crucial for road safety and traffic efficiency.
• Position falsification attacks threaten VANET integrity, compromising location-based services.
• Effective misbehavior detection frameworks are essential for secure VANET operation.

Purpose of the Study:

• To develop and optimize a stacked ensemble model for detecting position falsification attacks in VANETs.
• To enhance the accuracy of misbehavior detection through hyperparameter optimization.
• To improve the overall security and reliability of VANETs.

Main Methods:

• A stacked ensemble model was constructed using five base classifiers: KNN, ADA, ETC, RF, and XGBC.
• Logistic regression was employed as the meta-classifier to combine base model predictions.
• Artificial Bee Colony (ABC) optimization was utilized for hyperparameter tuning of the base classifiers.

Main Results:

• The optimized stacked ensemble model demonstrated superior performance in detecting position falsification attacks.
• The proposed method achieved higher accuracy compared to existing misbehavior detection techniques.
• Hyperparameter optimization using ABC significantly boosted the model's detection capabilities.

Conclusions:

• The optimized stacked ensemble model offers an effective solution for enhancing VANET security against position falsification.
• The methodology provides a robust framework for improving the accuracy and reliability of misbehavior detection systems.
• This research contributes to the development of safer and more efficient intelligent transportation systems.