Intelligent Electromagnetic Sensors for Non-Invasive Trojan Detection

Area of Science:

• Cybersecurity
• Sensor Technology
• Internet of Things (IoT)

Background:

• The proliferation of Internet of Things (IoT) devices necessitates robust security measures, especially for edge devices handling sensitive data wirelessly.
• Side-channel emissions, such as power consumption and electromagnetic (EM) emissions, offer a viable method for monitoring device behavior and detecting anomalies.
• Existing security solutions may not be suitable for energy-constrained, front-line applications at the extreme edge.

Purpose of the Study:

• To present a holistic self-testing approach for detecting security threats and malicious attacks directly at front-end sensors.
• To develop an integrated system using nanoscale EM sensing devices and an energy-efficient learning module for real-time threat detection.
• To enable on-chip, front-line security for energy-constrained wireless devices.

Main Methods:

• Incorporation of nanoscale electromagnetic (EM) sensing devices for monitoring side-channel emissions.
• Development of an energy-efficient learning module for analyzing sensor data and identifying threats.
• Deployment of intelligent EM sensors on power lines for detecting abnormal data activities.

Main Results:

• Successful implementation of a built-in threat detection system directly at the sensor level.
• Demonstration of effective detection of abnormal data activities without performance degradation.
• Achieved good energy efficiency, enabling on-chip detection for resource-limited devices.

Conclusions:

• The proposed holistic self-testing approach provides a viable solution for enhancing the security of edge devices in IoT.
• The intelligent EM sensors offer a low-power, high-efficiency method for rapid, front-line prediction of malicious attacks.
• This on-chip detection system is crucial for securing sensitive information transmitted wirelessly from the extreme edge.