Non-Intrusive Contact Respiratory Sensor for Vehicles

Area of Science:

• Biomedical Engineering
• Automotive Safety
• Wearable Sensors

Background:

• Monitoring driver physiological signals is crucial for enhancing vehicle safety.
• Non-intrusive methods for collecting respiratory data in vehicles are needed.
• Existing methods may be uncomfortable or impractical for continuous driver monitoring.

Purpose of the Study:

• To develop and evaluate a low-cost, non-intrusive system for collecting driver respiratory signals.
• To assess the accuracy of the system across different activities and driving conditions.
• To determine the optimal sensor configuration and signal processing for respiratory rate estimation.

Main Methods:

• Utilized a microcontroller with two piezoelectric sensors integrated into seat belt housings.
• Developed an iterative process for optimizing sensor housing design.
• Conducted tests in a driving simulator with participants performing various tasks (resting, deep breathing, driving, automated driving).
• Calculated breathing rates using a reference system, individual sensors, and sensor fusion, applying a 4-pole filter.

Main Results:

• A single chest-mounted sensor achieved an average absolute error of 0.92 min^-1 for respiratory rate.
• Error decreased significantly to 0.13 min^-1 during deep breathing exercises.
• Sensor fusion did not enhance system performance compared to a single sensor.
• A 4-pole filter with a 1 Hz cutoff frequency minimized errors across different conditions.

Conclusions:

• The proposed seat belt sensor system offers a robust and non-intrusive method for monitoring driver respiratory rate.
• Accurate respiratory rate assessment is achievable with a single sensor, simplifying the system.
• The technology shows potential for integration into vehicles to monitor driver well-being and alertness.