Heart Rate monitoring during physical exercise using wrist-type photoplethysmographic (PPG) signals

Heart rate monitoring using wrist-type using photoplethysmographic (PPG) signals during subjects´ intensive exercises is a challenging problem, since signals are strongly affected by motion artifacts caused by unexpected movements. This paper presents a method that uses both time and frequency characteristics of signals; using sparse signal reconstruction for high-resolution spectrum estimation. Experimental results on type data sets recorded from 12 subjects during fast running at peak speed of 15 km/hour. The results have a performance with the average absolute error being 1.80 beat per minute.