Malicious Activity Detection in Lightweight Wearable and IoT Devices Using Signal Stitching

Area of Science:

• Electrical Engineering
• Computer Engineering
• Cybersecurity

Background:

• Integrated circuit (IC) manufacturing involves complex global supply chains.
• Security risks arise from potential interference by rogue entities during IC fabrication and firmware development.
• Verifying manufactured devices for intended operations is crucial due to the economic infeasibility of controlling the entire supply chain.

Purpose of the Study:

• To present a novel detection technique for malicious activities originating from hardware or firmware Trojans in ICs.
• To address the challenge of verifying device integrity without access to trusted samples.
• To develop a method robust against environmental and device-to-device variations.

Main Methods:

• Repetitious side-channel sample collection from active devices.
• Time-domain stitching of collected samples.
• Frequency domain analysis, exploiting the low-frequency nature of Trojan activity for self-referencing.

Main Results:

• The proposed technique effectively detects malicious activities from hardware and firmware Trojans.
• Demonstrated effectiveness on wearable electronics prototypes and system-on-chip (SoC) devices.
• Achieved high detection coverage with a negligible monitoring time overhead of 0.8 seconds.

Conclusions:

• The developed technique provides a practical and efficient solution for IC security verification.
• Self-referencing in the frequency domain effectively mitigates variations, enabling reliable Trojan detection.
• The method offers a significant advancement in ensuring the security and integrity of manufactured electronic devices.