A Secure and Robust Machine Learning Model for Intrusion Detection in Internet of Vehicles

S SUNIL KUMAR
MTech, Department of Computer Science and Engineering
Vemu Institute of Technology, P-Kothakota,Chittoor
Andhra Pradesh – 517112, India
sunilkumar473@gmail.com

Mr. G. Lokesh
Assistant Professor, MTech Dept of Computer
Science of Engineering Vemu Institute of
Technology, P-Kothakota, Chittoor
Andhra Pradesh – 517112, India

Abstract: The research introduces a new Intrusion Detection System (IDS) which uses advanced machine learning methods to protect autonomous vehicles. The main purpose of the IDS system entails identifying cyberattacks which the system needs to categorize into different attack types that consist of DDoS Fuzzy Impersonation and standard Free traffic. The complete model development process uses the CAN-intrusion dataset which contains information about vehicle communication through its message and byte-level signal and target label data. The system uses ML techniques which include RF and Gradient Boosting and Adaboost and LSTM and CatBoost for security threat detection andprevention. The Random Forest and Decision Tree models  achieved 94% accuracy which makes them successful in detecting unauthorized access attempts to vehicle networks. The system achieves its maximum protection level through its complete execution of multiple security protocols. The new security enhancements will make smart vehicle systems more secure and reliable. The detection system develops as a scalable solution which protects smart vehicles from the growing number of cyber threats.
Keywords: RF, Gradient Boosting, Adaboost, LSTM, CatBoost classifiers, DDoS attack, Fuzzy attack, and Normal traffic patterns.