Bio-Inspired Approaches to Safety and Security in IoT-Enabled Cyber-Physical Systems

Area of Science:

• Cyber-Physical Systems (CPS) Security
• Hardware Security
• Bio-inspired Computing

Background:

• Attacks on Internet of Things (IoT) devices are increasing, yet hardware security in CPS remains underdeveloped.
• Existing security metrics often overlook hardware vulnerabilities, focusing on software, network, and cloud.
• Hardware Trojan attacks pose a significant threat to CPS integrity and functionality.

Purpose of the Study:

• To develop a novel bio-inspired approach for hardware Trojan detection in CPS.
• To leverage unsupervised learning and Spiking Neural Networks (SNNs) for identifying malicious electronic circuits.
• To enhance the security and trustworthiness of CPS in IoT applications.

Main Methods:

• Utilized a bio-inspired Spiking Neural Network (SNN) model, incorporating principles of glial cells for pattern identification.
• Developed a hardware Trojan detection circuit based on analyzing electronic circuit parameters as a fingerprint.
• Implemented a bio-inspired device-locking mechanism for a design-for-trust architecture.

Main Results:

• The SNN-based Trojan detection demonstrated stable firing patterns for normal device parameters and zero firing rate for Trojan presence.
• The proposed circuit exhibited resilience to various faults and attacks, including intentional and unintentional ones.
• Implementation on a Xilinx Artix-7 FPGA showed minimal hardware and power dissipation, suitable for resource-constrained environments.

Conclusions:

• The bio-inspired unsupervised learning approach offers an effective solution for hardware Trojan detection in CPS.
• The developed design-for-trust architecture enhances the security of CPS, particularly in IoT contexts.
• This work presents a new paradigm for secure CPS by integrating bio-inspired machine intelligence into hardware security.