This thesis focuses on monitoring movement and cardiac activity during a standardized yoga sequence using the Polar H10 chest strap. The aim of the thesis was to design and validate a procedure for detecting transitions between yoga postures based on accelerometer data and subsequently to analyse the cardiovascular response of the body in individual exercise segments.
The theoretical part describes the principles of acquisition and processing of accelerometric and cardiovascular data, particularly the ECG signal, RR intervals, heart rate, and heart rate variability. The practical part included the design of a yoga sequence, experimental data collection from 20 volunteers, and processing of signals obtained from the Polar H10 sensor. Two methods based on local variance and jerk of the accelerometer signal were compared for detecting transitions between postures. Detection accuracy was evaluated by comparison with manual annotations.
The results showed that both methods were able to detect transitions between yoga postures in close temporal proximity to the manual markers. The jerk-based method achieved a slightly lower overall mean absolute error; however, the difference between the methods was not statistically significant. Therefore, the local variance method was selected for the subsequent segmentation of cardiovascular data, as it provided a more stable detection pattern. Cardiovascular analysis was performed primarily using corrected RR intervals. Changes in heart rate, RMSSD, and recovery parameters after exercise were evaluated. Heart rate differed significantly between individual postures, with higher values observed mainly in postures with greater isometric and stabilization demands.
The proposed approach enabled the integration of accelerometer-based movement transition detection with segmented analysis of cardiac response. The results confirm the applicability of the Polar H10 sensor for orientational monitoring of yoga activities and for assessing differences between relaxation-oriented and more physically demanding parts of exercise. At the same time, the thesis highlights the need for RR interval quality control and cautious interpretation of HRV parameters in short movement segments.