Physical Noninvasive Attacks on Photoplethysmogram by Computer Controlled Blood Pressure Cuff

Area of Science:

• Biomedical Engineering
• Cybersecurity
• Wearable Technology

Background:

• Sensor data in welfare systems like insurance and healthcare enables personalized services.
• Altering sensor data poses a risk for unfair benefits, particularly through biometric manipulation.
• Photoplethysmogram (PPG) sensors in smartwatches are susceptible to such attacks.

Purpose of the Study:

• To propose and evaluate a noninvasive attack method to modify PPG sensor data.
• To investigate the feasibility of altering biometric information by targeting the wearer's body.
• To assess the impact of controlled blood volume manipulation on smartwatch heart rate readings.

Main Methods:

• A noninvasive attack was developed to manipulate PPG signals by controlling blood volume.
• Air pressure was applied to the upper arm to diminish pulse wave peaks.
• Experiments were conducted on seven subjects at rest and five post-exercise, using five smartwatch models and three pressure patterns.

Main Results:

• The proposed method successfully reduced the displayed heart rate on smartwatches.
• This effect was confirmed in both resting and post-exercise conditions.
• The attack effectively altered the sensor's measurement of true heart rate.

Conclusions:

• The study confirms the vulnerability of PPG-based heart rate monitoring to noninvasive physical attacks.
• This research highlights potential security risks in wearable health devices.
• Further investigation into robust anti-spoofing mechanisms for biometric sensors is warranted.