DMS Statistics and Data Science Seminar

Speaker: Dr. Chenang Liu, Oklahoma State University

Title: Data-Driven Anomaly Detection and Blockchain-Enabled Security Protection for Smart Manufacturing

Abstract: With the incorporation of the Internet of things (IoT) and information technologies, the environment of manufacturing become data-rich and cyber-enabled, which enables the rapid development of smart manufacturing. However, due to commonly existing the data imbalanced issue and cyber vulnerability, how to achieve accurate anomaly detection and effective cyber-security protection still remains challenging. To address these two critical issues, this research aims to develop methodologies that is capable of detecting the process anomalies accurately and preventing the potential cyber-physical attacks effectively. Toward this goal, an adversarial learning-based approach termed augmented time regularized generative adversarial network (ATR-GAN) is proposed to handle the highly imbalanced sensor data for anomaly detection. Then a blockchain-enabled method is also developed to prevent the common cyber-physical attacks in manufacturing. Furthermore, real-world case studies based on additive manufacturing was also conducted to demonstrate the effective and potential of the proposed methods.

Location: https://auburn.zoom.us/j/82501343299