Estimation of Sleep Quality Based on HRV, EMG, and EEG Parameters with K-Nearest Neighbor Method

Fatigue after sleeping can be a significant indicator that a person has poor sleep quality. The K-Nearest Neighbor (KNN) algorithm method was used in this study to objectively identify fatigue in subjects after sleeping, based on the body condition parameters for determining the sleep quality. This study proposes a comprehensive evaluation of sleep quality to identify fatigue after a sleep period, using a combination of electrocardiography (ECG), electromyography (EMG), and electroencephalography (EEG). Physiological signal analysis of the body is performed under a variety of conditions. Several parameters are obtained by conducting Heart Rate Variability (HRV) analysis of the ECG signal in both the time and frequency domains: Low-High Frequency (LF/HF) ratio, the proportion of NN50 divided by the total number of NN (R-R) intervals (PNN50), and Root Mean Square of Successive Differences (RMSSD). The EMG and EEG signals were subjected to spectral analysis to extract the mean power frequency (MPF). Ground truth in the dataset for the KNN is carried out by evaluating the Fatigue Severity Scale (FSS) and MPF values from the EEG signal. Tests were carried out on ten subjects with healthy conditions after sleeping. Subjective testing was also applied using the Groningen Sleep Quality Scale (GSQS) assessment. The process of evaluating the results of the GSQS questionnaire with an objective review of fatigue levels provided a similar level of sleep quality, as much as 72% of accuracy.