Improving Sleep Apnea Detection Performance by Reducing Artifacts in Respiratory Signals Using Digital Filters

IMPROVING SLEEP APNEA DETECTION PERFORMANCE BY REDUCING ARTIFACTS IN RESPIRATORY SIGNALS USING DIGITAL FILTERS This study addresses the challenge of accurately detecting sleep apnea due to the presence of artifacts or noise in respiratory signals, which affects diagnostic reliability. The primary goal of this research is to design a device that enhances apnea detection by reducing noise and artifacts in respiratory signals using digital filtering, while also providing automatic respiratory support and monitoring respiration rate. The study contributes by developing a more accurate and efficient sleep apnea detection device, utilizing a piezoelectric sensor to measure the respiration rate and applying a Chebyshev digital filter to improve signal clarity. The device uses an Arduino AT-Mega328 microcontroller for real-time signal processing and apnea monitoring. The research design involves building the hardware system, programming the software, testing the device with participants, collecting data, and validating the effectiveness of the digital filter. The results indicate that the device successfully detects apnea events with precision, significantly reduces noise, and provides accurate respiration rate monitoring, with errors recorded as 0.58%, 0%, 0%, 0.58%, 0.01%, 0.42%, 0.15%, 0.43%, 0.05%, and 1.06%. The largest error rate is 1.06%, the smallest is 0%, and the average error is 0.32%. The apnea time error rate results show percentages of 0.42%, 0.35%, 0.35%, 0.44%, and 0.27%. The highest error rate is 0.44%, the lowest is 0.27%, with an average error rate of 0.36%. The conclusion demonstrates the effectiveness of the device in improving sleep apnea monitoring, and its implications suggest that this device could become a reliable and cost-effective solution for patients, enhancing healthcare outcomes and quality of life..