Design Of Fetal Movemnet Signal Acquisition System And FM Recognition Algorithm

Perinatal monitoring technology has great significance to ensure the safety of mother and fetus.Fetal movement monitoring is an important index for assessment of fetal physical development and neurological development during the perinatal period.Reduced fetal movement may lead to adverse outcomes,such as fetal congenital malformations,preterm birth,cerebral palsy,poor mental retardation,low birth weight,ketogenic hypoglycemia,fetal death,early neonatal death,increased risk of planing of women,and so on.The traditional pregnant women self-counting has low accuracy,and ultrasound fetal movement monitoring has potential harm for futus.At present,the main methods of fetal movement detection are ultrasonic detection,acceleration sensor detection,pressure sensor detection and so on.Ultrasound detection is mainly used in hospitals,because ultrasound instrument can not carry out long time fetal movement detection,and needs professional doctor operation.It can not be applied at home.There are some difficulties in the practical application of other fetal movement detection methods.This study focuses on these issues from the following aspects:1.In this study,a fetal movement sensor containing 4 piezoelectric thin films was designed to collect the fetal movement signal.The piezoelectric thin films is soft and light.It can detect the fetal movement signal in different position.and does not need to be powered when it is used.This guarantees the safety and comfort of pregnant women.2.In this study,a portable fetal movemnet acquisition system which can be used for a long time was designed for using at home.The output of the fetal movement sensor was processed and collected to the PC side via Bluetooth.The PC-side control storage analysis software was designed.3.Aiming at the characteristics of noise and discrete values of the original signal collected,the IIR high-pass filter is used for denoising.The discrete value is removed by medain absolute deviation method.The high signal-to-noise ratio is obtained to identify the fetal movement signals.4.In this study,the time-frequency analysis method of short-time Fourier transform and pseudo-Weignewilli distribution were used to analyze the fetal movement signal.The 12 characteristic parameters of the fetal movement signal were extracted by sliding window.The effect of each characteristic parameter on fetal movement detection was evaluated.Finally,the maximum,minimum,RMS,variance,deviation,kuriosis,energy concentration and renyi entropy were selected.The fetal movement recognition model based on automatic threshold was established to identify fetal movement.The model was tested.The ture detection rate of fetal movement was 93.0%,and the positive predicted value was 75.9%.This study achieved good results.