An Improved Auto-Correlation Algorithm and its Application in Fetal Heart Rate Detection

Through the research of the auto-correlation and its improved algorithms in pitch estimation, this paper presents an auto-correlation main peak locating method based on the analysis of peaks value approximation and measure range features, and establishes an adaptive algorithm model using variable sampling rate and segmental Circular Average Magnitude Difference Function (CAMDF). This model is successfully applied in Ultrasonic Doppler Fetal Heart Rate (FHR) signal analysis and processing, effectively eliminating the misjudgment especially due to doubling and halving in the detection of FHR. The algorithm passes through the simulation with Matlab and is successfully transplanted to a portable fetal monitor designed with AVR processor. Experimental and clinical results indicate that the FHR monitor using above algorithm model has good detection performance and strong practicality.