IMPROVE: Intelligent Machine Learning based Portable, Reliable and Optimal VErification System for Future Vehicles

Abstract

The technological progress over the past decade has revolutionized the transportation domain. Autonomous and semi-autonomous vehicles have now gained the global spotlight for facilitating personal transportation with minimal manual intervention. The digitization of this industry has been accompanied by significant security challenges in terms of ensuring reliable transmission and robust communication networks which are critical for the proper functioning of the smart vehicle. The CAN bus architecture responsible to establishing connectivity within the various vital components of the car’s internal architecture is a prime target for intrusions. Secure connections must also be established between the vehicle and external devices such as smartphones for enhancing the travel experience. Hence a complete security intrusion detection framework for self-driving cars is of dire need. This article introduces an Intelligent Machine Learning based Portable, Reliable and Optimal VErification System (IMPROVE) for Future Vehicles that aims to provide a viable solution to resist vehicular cyberattacks both on the internal network of the vehicle and the vehicle to device network established. The proposed framework is twofold in nature- The initial module focusses on ensuring Controller Area Network (CAN) security through machine learning modelling for intrusion detection. The second module is oriented towards utilizing data analysis to detect and block malicious behaviour on networks established with external/internal devices.