ECG baseline wander removal and noise suppression analysis in an embedded platform

ECG signal processing in an embedded platform is a challenge which has to deal with several issues. One of the commonest problems in ECG signal processing is baseline wander removal and noise suppression, which determine posterior signal process. In this report, two filtering techniques are presented and implemented to work on a Shimmer platform. Baseline wander removal based on cubic splines and morphological filtering are evaluated to check whether are suitable for realtime execution. The use of cubic splines is made to estimate the baseline wander in an ECG signal and then substract it from the input dataset to remove the baseline wander. Morphological operators are useful for signal processing and noise suppression. These techniques have been implemented and tested by a wavelet-based delineation algorithm and results are provided for comparison purposes. The project goal is to develop an implementation for baseline wander removal and noise suppression to be executed on an embedded platform, meeting its specific hardware constraints, and leaving room for posterior signal processing. This would allow to design a Wireless Body Sensor Network to support non-ambulatory healthcare.