Physics Meets AI: Hardware Authentication for Trustworthy Automotive Systems - Colloquium Presentation by: Dr. Hafiz Malik

Abstract: The increasing reliance on autonomous and advanced driver-assistance systems (ADAS) has introduced critical security vulnerabilities, including hardware counterfeiting and supply chain attacks. To establish trust in these systems, we present a novel "Physics Meets AI" framework for hardware authentication, designed to secure in-vehicle electronic components such as Electronic Control Units (ECUs) and ADAS. Our approach leverages the inherent uniqueness of electronic devices, where intrinsic analog imperfections in signal generation or processing paths create a device-specific signature, or "hardware fingerprint." By extracting key features from these physical-layer signals and applying machine learning classification, we effectively link data to its verified hardware origin, forming a foundational trust layer. This talk will showcase a physics-based electronic device authentication use case, demonstrating an optimized feature space and a computationally efficient algorithm suitable for real-time deployment in resource-constrained automotive environments. Our technique achieves exceptional classification accuracy and maintains performance across diverse operational conditions, including variations in sensor hardware and sampling rates. As a non-crypto-based solution, it provides robust defense against hardware manipulation, enabling the development of truly trustworthy automotive systems.

Biography: Hafiz Malik is a Professor and College of Engineering and Computer Science (CECS) Associate Dean of Graduate Education and Research at the University of Michigan (UM)–Dearborn. His research in the areas of automotive cybersecurity, cyber-physical system security, sensor security, deepfakes, multimedia forensics, steganography and steganalysis, information hiding, pattern recognition, and information fusion is funded by the National Science Foundation, the National Academies, Google, Ford Motor Company, Marelli, Inc., and other agencies. He is a recipient of the UM-Dearborn Chancellor's Inclusive Excellence Fellows 2022, UM-Dearborn 2022 Distinguished Research Award, and CECS 2020 Excellence in Research Award. He is the founder and CTO of Media Shield Inc. and Co-Founder and Chief Scientist at Tensor Machines, Inc. He is also a founding member and COO of the Global Foundation for Cyber Studies and Research, a director and founding member of the Director of Cybersecurity Center for Research, Education, and Outreach at the UM-Dearborn, and a member of the leadership circle for the Dearborn Artificial Intelligence Research Center at the UM-Dearborn. He was a member of the Scientific and Industrial Advisory Board of the National Center of Cyber Security, Pakistan, from 2020 to 2023. He has been a member of the MCity Working Group on Cybersecurity since 2015.