Machine Learning-based Secure Data Communication Framework For Autonomous Electric Vehicles

Autonomous Electric Vehicles (AEVs), being an innovation in the transportation sector, are powered by electricity and technology to operate independently. However, to achieve self-driven abilities, AEVs depend on wireless communication networks and software systems, which increase susceptibility to cyber-attacks. Notably, such vulnerability can undermine passenger safety and the entire transportation network. This research examines Logistic Regression and Decision Tree classification Machine Learning (ML) techniques, focusing on enhancing AEV security against cyber threats. As part of the research, the following cyber attack scenarios were considered, and the effectiveness of ML techniques in real-time threat detection was reviewed. This study aims to develop stringent intrusion detection systems to protect AEVs in a practical scenario. Moreover, a system model and a problem formulation are provided to address cybersecurity challenges in AEVs. Through features like duration, protocol type, service, and others, this research apply the Gini index and gain ratio approaches for feature selection. Utilizing supervised ML algorithms, this research forecasts threatening actions and evaluates how acutely and swiftly they can be identified. The results of our research add to the general idea of cybersecurity in AEVs development processes. Continuous learning and monitoring are crucial in the context of unpredictable surroundings.