Heart Rate Estimation Using FMCW Radar: A Two-Stage Method Evaluated for In-Vehicle Applications

Area of Science:

• Automotive Safety
• Biomedical Engineering
• Sensor Technology

Background:

• Driver monitoring is vital for vehicle safety, with human factors causing most accidents.
• Contactless heart rate (HR) measurement in vehicles is challenging due to motion and environmental factors.

Purpose of the Study:

• To develop and validate a novel two-stage method for accurate, contactless driver HR estimation using 60 GHz FMCW radar.
• To improve driver state assessment in real-time for advanced vehicle safety systems.

Main Methods:

• A two-stage HR estimation approach combining discrete wavelet transform (DWT), autoregressive (AR) spectral analysis, and relevance vector machine (RVM).
• Sequential Kalman filtering (SKF) for stabilizing HR estimates.
• Implementation with an Infineon BGT60TR13C radar module in a vehicle's sun visor.

Main Results:

• Robust and accurate HR estimations achieved, comparable to commercial wearable devices.
• Successful validation against a Polar H10 chest strap during real-world driving conditions.
• Demonstrated effectiveness with short measurement windows (5s) and under varied lighting/clothing conditions.

Conclusions:

• The proposed radar-based system offers a viable solution for contactless driver HR monitoring in dynamic automotive environments.
• This method advances biomimetics by applying biological vital sign perception principles to automotive sensorics.
• Future work includes sensor fusion with visual methods and heart rate variability (HRV) estimation for enhanced driver monitoring systems (DMSs).