Security Challenges and Solutions for Autonomous Vehicles and Drones in the AI Age
Venkatraman Balasubramanian, Moayad Aloqaily, Mohsen Guizani, Bassem Ouni · IEEE Internet of Things Magazine · 2025
The rapid integration of AI into autonomous systems like vehicles and drones raises significant security concerns. This paper focuses on analyzing the vulnerabilities and potential threats faced by autonomous systems in the context of AI. We emphasize the importance of implementing robust security measures to ensure the safe operation and protection of these systems, particularly when applied in Niche Critical Infrastructure (CI). Furthermore, we conduct a prototypical experiment to demonstrate the effectiveness of anomaly detection in enhancing the security of autonomous systems through the application of reinforcement learning techniques. By leveraging anomaly detection, autonomous systems can proactively identify and respond to abnormal behaviors, thereby safeguarding CI from potential risks and ensuring reliable and secure operations. Our research highlights the critical role of anomaly detection in safe-guarding autonomous systems and underscores the need for further advancements in security solutions to address emerging threats in this evolving landscape. By adopting proactive security measures and integrating anomaly detection with reinforcement learning, we can enhance the resilience and trustworthiness of AI-driven autonomous systems, promoting their widespread adoption in various domains while minimizing potential risks to CI.