ECG/PPG integer signal processing for a ubiquitous health monitoring system

Abstract

A compact ubiquitous-health monitor operated by single 8-bit microcontroller was made. An integer signal processing algorithm for this microcontroller was developed and digital filtering of ECG (electrocardiogram) and PPG (photoplethysmogram) was performed. Rounding-off errors due to integer operation was solved by increasing the number of effective integer digits during CPU operation; digital filter coefficients and data expressed in decimal points were multiplied by a certain number and converted into integers. After filter operation, the actual values were retrieved by dividing with the same number and selecting available highest bits. Our results showed comparable accuracies to those computed by a commercial software. Compared with a floating-point calculation by the same microcontroller, the computation speed became faster by 1.45 ∼ 2.0 times depending on various digital filtering cases. Our algorithm was successfully tested for remote health monitoring with multiple users. If our algorithm were not used, our health monitor should have used additional microcontrollers or DSP chip. The proposed algorithm reduced the size and cost of our health monitor substantially.