Hybrid autonomous guided vehicle (AGV) and drone systems represent a significant advancement in industrial automation, yet their integrated circuits (ICs) face critical cybersecurity vulnerabilities. Their interconnected IC components create expanded attack surfaces vulnerable to sophisticated cyber-attacks that enable covert remote control. This research aims to develop and validate a machine learning-enabled (ML-enabled) detection system for identifying and preventing unauthorized access attempts targeting the interconnected IC components of hybrid AGV-drone platforms. Our methodology implemented real-time IC behavioral monitoring using distributed sensor networks across both AGV and drone platforms. The system employs a multi-layer detection approach, combining signal analysis and pattern recognition with machine learning algorithms to identify security breaches. The implemented system achieved 95% accuracy in detecting unauthorized access attempts, with response times averaging under 10ms for rapid threat mitigation. False positive rates remained below 2% during extensive testing across different environmental conditions. The system successfully identified and blocked 98% of simulated remote manipulation attempts targeting both platforms. Cross-platform threat detection showed 96% accuracy in identifying attacks exploiting the system's interconnected nature. We recommend implementing this ML-enabled security framework as a standardized component in hybrid AGV- drone systems, with regular updates to address evolving attack patterns.

Keywords: Machine Learning, Anomaly Detection, Integrated Circuit Security, Remote Attack Detection, Autonomous Robotics, Industrial Security, Behavioral Analysis, Cybersecurity, Neural Networks, Real-time Monitoring