Motion noise cancellation in seismocardiographic monitoring of moving subjects

This paper presents a novel method of extracting seismocardiographic data from moving adult subjects using chest-worn wireless MEMS accelerometers. A DSP system including a normalized LMS adaptive filter is designed and tested in MATLAB. Data results from 11 subjects indicate a detection rate of 97.33% and a high correlation between SCG and ECG signals simultaneously acquired. The results suggest that adaptive filtering could be a relevant technique for this purpose.