Robotic Cloud Automation-Enabled Attack Detection and Command Verification Using Attention-Based RNNs, ConvLSTM, and Bayesian Networks
Keywords:
Automated robotics, security in the cloud, artificial intelligence, detecting intrusions, verifying commands, recurrent neural networks, ConvLSTM, networks based on Bayesian statistics,, identifying anomalies, information securityAbstract
Background Information: The emergence of robotic cloud automation has brought about fresh cybersecurity hurdles, particularly in protecting communication and control systems from cyber threats. It is crucial to guarantee strong intrusion detection and verify commands effectively. Objectives: Create an AI framework by combining deep learning and probabilistic models to improve intrusion detection and command verification in cloud-based robotic systems. Methods: The system combines Attention-Based RNN, ConvLSTM, and Bayesian Networks to identify abnormalities and authenticate instructions, utilizing temporal and spatial data for instant threat identification. Results: The combined model demonstrates a high level of accuracy (96.4%) along with a low rate of false positives (1.5%), which improves the overall security effectiveness. Conclusion: In summary, this method successfully improves the security of cloud-based robots, providing a scalable and efficient way to combat cyber threats in rapidly changing environments.
