This paper proposes new heart rate estimation method for biological data from sleep monitor sensor toward estimating sleep stage accurately. Concretely, we employed two heart rate estimation methods, and integrated the two methods as weak estimator. One of the two methods calculates power spectrum from the biological data by FFT, and selects the frequency with maximum spectrum as heart rate (HR). The other calculates power spectrum as a same manner of the former method, and selects the frequency which indicates the half size of all power spectrum as HR. To validate the effectiveness of EHEM, this paper applies EHEM to pressure data from sleep monitor sensor. From the result, EHEM can extract HR accurately, and prevent from outliers generated by HEM-FFT. We are going to research (1) what method gives good influence to EHEM, and (2) how to integrate the HRs extracted from the methods.
題目: Ensemble Heart Rate Extraction Method for Biological Data from Water Pressure Sensor
著者: Fumito Uwano and Keiki Takadama
誌名: Proceedings of AAAI Spring Symposium 2018
詳細: Stanford University (San Francisco), USA, May 2018