An extended Kalman filter for fetal heart location estimation during Doppler-based heart rate monitoring

Fetal heart rate (fHR) monitoring using the Doppler ultrasound (US) is a standard clinical practice for assessing fetal well-being before and during labor. For continuous fHR measurements, the US transducer is positioned on the maternal abdomen using a flexible belt. Due to fetal movement, the relative fetal heart location (fHL) with respect to the US transducer can change, leading to frequent periods of signal loss hampering the clinical assessment of fetal health. Consequently, the clinical staff has to repeatedly reposition the US transducer--a cumbersome task affecting clinical workflow. We propose a method to estimate the fHL during fHR monitoring to support clinicians in efficiently repositioning the US transducer. Unlike typical US transducers, which do not provide any information on the spatial fHL, we exploit the fact that multiple transducer elements are present in the array aperture of the US transducer. We developed a novel model that relates the measured Doppler power in the individual transducer elements to the fHL and use it within the probabilistic framework of an extended Kalman filter (EKF). The performance of the EKF algorithm was evaluated in simulations and in in vitro experiments using a dedicated setup of a beating fetal heart. Both simulations and in vitro experiments showed that the fHL can be determined with an accuracy of 4 mm. Furthermore, we demonstrate that when the fetal heart is drifting out of the US beam, the EKF algorithm accurately estimates the fHL up to a radial distance of $34$ mm.