Interbeat Interval Filtering

Interbeat intervals (IBIs) carry heart rate and heartrate variability information. When the IBIs are correct, this information can be extracted through simple statistics such as mean/median/standard deviation. However, more robust approaches are needed especially in ambulatory conditions where the underlying signal can be noisy. In this paper, we propose a lightweight filter that (i) tracks the time-varying distribution of interbeat intervals, (ii) is robust to outliers. The tracked distribution summarizes relevant information of interest and alleviates the need for further analysis. We demonstrate numerically that the tracked distribution can provide more accurate estimates than a two-stage algorithm that first fixes the IBIs before computing the statistics of interest.