Enhancing Contactless Respiratory Rate Measurement Accuracy: Integration of 24GHz FMCW Radar and XGBoost Machine Learning

Advancements in non-contact vital sign monitoring are crucial for enhancing patient measurements' accuracy and overall patient experiences. This research explores the integration of 24GHz Frequency-Modulated Continuous-Wave (FMCW) radar with the XGBoost machine learning algorithm to improve the detection of respiratory rate (RR). This innovative approach offers a promising alternative to traditional contact-based methods. The study utilizes FMCW radar to detect respiratory motion, while signal patterns are analyzed using XGBoost to ensure accuracy across various healthcare environments. The method involves collecting signals, pre-processing to remove noise and irrelevant data, and extracting features to be analyzed by the XGBoost algorithm. The collected dataset, which includes controlled and randomized respiratory rates from a diverse subject pool, establishes a solid basis for the algorithm's training and validation, ensuring extensive adaptability and precision. Empirical results show that XGBoost surpasses other machine learning models' accuracy and reliability. Importantly, this method significantly reduces error margins compared to established benchmarks, leading to substantial improvements in RR measurement. The implications of this study are wide-ranging, indicating that such a system could significantly enhance patient care standards by providing continuous, accurate, and non-intrusive monitoring, especially in settings where traditional methods are impractical or uncomfortable. Future research should aim to refine the system's real-world applicability, assess long-term reliability, and optimize the technology for integration into existing healthcare frameworks, thereby further transforming the landscape of patient monitoring technologies.