Artificial Intelligence for Automotive Security: How to Support Developers in Automotive Solutions

With newer and more advanced technologies in the automotive industry come complex and sinister threats. To thwart attacks targeted at a vehicle's Controller Area Network (CAN), developing new and increasingly sophisticated algorithms exploiting Machine Learning and Deep Learning is not enough. Thus, where some of these techniques are impractical to embed into vehicles mounting Electronic Control Units (ECUs) with limited computational power, others may even require high prediction time, a critical factor in Cyber-Physical Systems (CPS) like modern vehicles. Moreover, help and support for developers during the automotive development life cycle are unorganized and inconsistent. This paper presents a study of Traditional and Quantum Machine Learning algorithms to improve the identification of in-vehicle attacks and redefine the security life cycle in this context serving as a baseline to define a vision model based on Artificial Intelligence that supports developers' decisions to integrate concrete design methods in the automotive development life-cycle.