A Circuit-Level Solution for Secure Temperature Sensor.

Area of Science:

• Electrical Engineering
• Computer Science
• Cybersecurity

Background:

• Internet-connected sensors, including temperature sensors, face increasing security and integrity concerns.
• Existing system-level security measures for sensors are inefficient, causing high overhead in delay and power consumption.
• Low-end sensors lack built-in defense mechanisms, necessitating novel security architectures.

Purpose of the Study:

• To propose a secure architecture for temperature sensors that enhances resilience against intentional and unintentional attacks.
• To develop a detection system that operates at both the transducer and signal conditioning levels.
• To improve the efficiency of anomaly detection compared to traditional system-level countermeasures.

Main Methods:

• Developed a secure temperature sensor architecture incorporating a transducer and a signal conditioning unit.
• Implemented statistical analysis for sensor data estimation and residual signal generation for anomaly detection.
• Utilized complementary current-temperature characteristics to create a constant current reference for transducer-level attack detection.

Main Results:

• The proposed architecture successfully detected under-powering attacks and analog Trojans via signal vibration analysis in the constant current reference.
• Anomaly detection at the signal conditioning unit identified anomalies using the generated residual signal.
• The integrated detection system demonstrated a high attack detection rate of 97.73%.

Conclusions:

• The proposed secure temperature sensor architecture effectively enhances attack resilience through multi-level detection mechanisms.
• This approach offers a more efficient and targeted solution for sensor security compared to generic system-level defenses.
• The system provides robust protection against a wide range of intentional and unintentional security threats.