Wavelet based data analysis for implantable pulse oximetric sensors.

Cardiovascular data recording by implantable sensor modules exhibits a number of advantages over extra-corporeal standard approaches. Implantable sensors feature their benefits in particular for high risk patients suffering from chronic heart diseases, because diagnosis can be combined with therapy in a closed loop system. Nevertheless, the measured photoplethysmographic signals reveal different kinds of noise and artifacts. There are several parametric and non-parametric mathematical techniques that try to achieve optimality and generality in estimating the actual signal out of its noisy representation. The determination of blood oxygen saturation and pulse transit time requires one of these mathematical techniques for gaining the exact position and magnitude of maxima and minima in the photoplethysmograph. A robust wavelet algorithm resolves the difficulties arising from physiological data.