Using Adaptive FIR and IIR Filtering to Process Fetal Electrocardiogram Signals
Open Access
- Author:
- Sidehamer, Adam Henry
- Area of Honors:
- Electrical Engineering
- Degree:
- Bachelor of Science
- Document Type:
- Thesis
- Thesis Supervisors:
- William Kenneth Jenkins, Thesis Supervisor
Dr. Timothy Joseph Kane, Thesis Honors Advisor - Keywords:
- fetal electrocardiography
adaptive signal processing
FIR filters
IIR filters
linear prediction
adaptive noise cancellation - Abstract:
- Fetal electrocardiography can provide useful data for monitoring a fetus’s cardio health and diagnosing a wide range of heart disorders. The most popular approach is the non-invasive method, which is useful for the second half of pregnancy and safer for the fetus. Unfortunately, this method also requires extensive signal processing to extract the fetal heartbeat from the maternal interference that is also picked up by the fetal electrocardiogram. One of the leading methods in extracting the fetal heartbeat consists of sequentially applying the following three processes: • Adaptive linear prediction (LPC), • Adaptive noise cancellation (ANC), and • Comb filtering. There are numerous ways of implementing each process, but some are more efficient and/or accurate than others. This paper covers the time-domain implementation of a direct form IIR filter in place of the FIR filter in the LPC stage of the clean-up process. Also, FIR filtering is examined closely in the time and frequency domains of the LPC and ANC stages. All possible combinations are executed and compared for two different abdominal signals. The process involves writing the algorithms with MATLAB and testing them with real and modified synthetic data from the PhysioNet archives.