An intelligent IDS using bagging based fuzzy CNN for secured communication in vehicular networks

Area of Science:

• Cyber-physical systems
• Intelligent transportation systems
• Network security

Background:

• The Internet of Vehicles (IoV) is crucial for intelligent transportation systems but faces increasing cyber-physical attacks.
• Existing Intrusion Detection Systems (IDS) often lack sufficient accuracy and have high false positive rates.
• Effective security solutions are needed to mitigate cyber threats in vehicular networks.

Purpose of the Study:

• To propose an efficient feature selection algorithm and a novel classification algorithm for enhanced intrusion detection in IoV.
• To improve the accuracy of attack identification while minimizing false positives in vehicular cyber-physical systems.
• To address the limitations of existing IDSs in detecting sophisticated cyber-physical attacks.

Main Methods:

• Developed a Weightage and Ranking Based Feature Selection Algorithm for efficient feature selection.
• Proposed a Bagging based Fuzzy Convolutional Neural Network (FCNN) classification algorithm with Adam optimizer.
• Integrated fuzzy inference within a deep convolutional neural network classifier for improved attack identification.

Main Results:

• The proposed IDS demonstrated enhanced detection accuracy compared to existing methods.
• A significant reduction in the false positive rate was observed.
• The system proved effective on both benchmark and network trace datasets.

Conclusions:

• The proposed Weightage and Ranking Based Feature Selection Algorithm and Bagging-based FCNN offer a more effective solution for IoV security.
• The developed IDS significantly improves intrusion detection performance by enhancing accuracy and reducing false positives.
• This research contributes to more robust security for intelligent transportation systems against cyber-physical attacks.