An Adaptively Parameterized Algorithm Estimating Respiratory Rate from a Passive Wearable RFID Smart Garment

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Signal Processing

Background:

• Wired respiratory rate sensors present limitations including patient tethering, potential body obscuration, and issues with accuracy or invasiveness.
• The Bellyband, a smart garment sensor utilizing wireless, passive Radio Frequency Identification (RFID), offers a less invasive solution but faces signal processing challenges due to noisy, quantized data.

Purpose of the Study:

• To develop and validate an advanced algorithm for estimating respiratory rate from the Bellyband sensor's bio-signals.
• To improve the accuracy, speed, and real-time applicability of respiratory rate detection from the Bellyband compared to existing methods.

Main Methods:

• An adaptively parameterized Savitzky-Golay (SG) filter was employed to smooth the noisy signals from the Bellyband.
• The algorithm's adaptive parameterization allows for effective processing across a range of respiratory frequencies, including sharp changes.
• Data from 25 trials, collected using an off-the-shelf respiratory monitor and metronome-synchronized breathing, were used for testing.

Main Results:

• The new algorithm demonstrated a significant improvement, being three times faster and three times more accurate than the previous Bellyband algorithm.
• Respiratory rate estimates achieved an average Root Mean Square Error (RMSE) of 4.1 breaths per minute (BPM) against ground truth.
• Preliminary findings indicate potential for accuracy comparable to or exceeding established Electrocardiogram (ECG) or Impedance Plethysmography (IP) based methods for non-ventilated patients.

Conclusions:

• The developed algorithm effectively processes noisy RFID signals from the Bellyband for accurate respiratory rate estimation.
• This advancement offers a promising, real-time, wireless, and less invasive solution for respiratory monitoring.
• The algorithm shows potential to rival current standards in respiratory rate detection for non-ventilated patients.