AN AUTO-NORMALIZING ZERO-CROSSING DETECTION SYSTEM FOR WEARABLE GARMENT-BASED RESPIRATORY RATE MONITORING

Monitoring respiratory rates is critical in understanding and predicting medical events but has proven difficult to reliably and noninvasively measure. Previously, a garment was developed using conductive thread-based sensors with the ability to output an electrical signal representing the respiration of the wearer due to a variable resistance when stretched. Detection was performed using a high-gain instrumentation amplifier and zero-crossing rate detection, which has significant advantages over frequency-based detection methods, but the amplifier needed to be carefully manually calibrated for each user and was susceptible to errors due to drift. A new design presented in this paper uses a narrow-bandwidth high-gain differentiating amplifier to automatically calibrate the system and a Schmitt-trigger based zero-crossing rate detection algorithm for noise rejection.. The system is tested on a sample of athletic wear fabric with a conductive thread sensor controlled by a variable-speed stepper motor to simulate motion due to breathing. The system is shown to have accurate detection ability at three example respiratory rates of 12.2, 13.9, and 14.6 breaths per minute.