Designing Robust AI Systems for Autonomous Vehicles Against Adversarial Attacks
Thirunavukkarasu Mani, Shweta Saxena, Shalini, Ashwini Ranjit Nawadkar, N. Kalidas, Abhinav Rathour · 2024
More people are purchasing autonomous vehicles (AVs), which raises concerns about how simple it would be for bad guys to attack them. AI-powered devices must be trustworthy to keep people and automobiles safe as they improve. This article suggests we need AI systems to discover and reduce self-driving vehicle threats immediately. You require real-time monitoring, outlier detection, and good training techniques for this position. Strong training approaches safeguard the AI model from unintended modifications with multiple datasets. Strange AV activity or patterns may be detected using anomaly detection, allowing you to take immediate safety measures. Advanced deep learning algorithms are crucial for real-time threat detection and reduction, according to the research. Because of its surveillance and analysis tools, the AI system can instantly adapt to changing events and eliminate potential hazards before they become safety risks. To ensure all antivirus businesses deliver top-notch protection, this article advocates for consistent standards and government monitoring. Risk analyses, system upgrades, and security all work to protect AV network data and communication connections. The recommended basic architecture strengthens self-driving vehicle AI systems to manage smart threats from other drivers and protect passengers and onlookers. This work is crucial to ensuring that antivirus products remain safe and trustworthy as digital threats and vulnerabilities develop.