BReML: A Breathing Rate Estimator Using Wi-Fi Channel State Information and Machine Learning

Breathing rate is one vital sign that might help identifying pathological conditions by its monitoring. This paper presents a novel breathing rate estimator that combines conventional Channel State Information (CSI) approaches with machine learning classifiers to provide a breathing rate estimation in a controlled environment. Results show that by extracting time and frequency domain features of CSI amplitude, as well as using a first estimation obtained with Fast Fourier Transform as a feature for feeding machine learning classifiers, the breathing rate can be accurately estimated.