Ensuring Hardware Reliability and Security Using Terahertz <scp>VLSI</scp> ‐Based Tree Hierarchical Deep Convolutional Neural Network Testing Technology

ABSTRACT Hardware cyber security is a major issue in very large integrated circuits (ICs). The reliability and security of the system are vulnerable to counterfeit, forged, or defective ICs. The increasing complexity of digital and mixed‐signal systems makes it difficult to develop effective techniques for evaluating and validating the reliability of ICs. In this manuscript, ensuring hardware reliability and security using terahertz very‐large‐scale‐integration (VLSI)–based tree hierarchical deep convolutional neural network testing technology (EHRS‐TZ‐VLSI‐THCNN) is proposed. A focused terahertz laser is used to scan a radio frequency integrated circuit (RFIC), and the response on a single pin is evaluated to generate two dimensional (2D) pictures. The secure picture data set is created by a data augmentation technique to train the proposed algorithm. The proposed EHRS‐TZ‐VLSI‐THCNN approach achieves 25.26%, 31.74%, and 25.21% greater precision and 21.52%, 24.244%, and 25.65% better accuracy when compared to the existing models: AI‐powered terahertz VLSI testing technology to ensure hardware safety (CNN‐PT‐VLSI‐EHS), AI powered terahertz (THz) VLSI testing technology (PNDT‐PT‐VLSI‐EHS), and AI powered THz testing technology for ensuring hardware cybersecurity (AI‐PT‐VLSI‐EHS).