This research focuses its attention on the study of a cross-coupled adaptive noise cancellation algorithm that is suitable for applications in which the primary and noise inputs may be closely spaced, thus a pure noise reference is not available for effective noise cancellation from the primary signal. The structure of this new algorithm is based on two Least-Mean Squared (LMS) delay line cancellers with cross-coupled feedback. The PI's overall objective is to determine the feasibility of this novel cross-coupled adaptive algorithm to improve the signal-to-noise ratio of data composed of recurring pulses in noise for systems in which a pure noise reference input can not be found, i.e., in the problem of extracting the fetal electrocardiographic (FECG) signals embedded in muscle noise recorded from the pregnant maternal abdomen. He hopes to show that this LMS cross-coupled canceller algorithm, which does not require a pure noise reference input, effectively cancels maternal muscle noise from recorded fetal ECG data without distortion of the FECG signal, thus, providing an enhanced fetal signal for subsequent analysis or detection. The evaluation of this cross-coupled cancellation algorithm for the real-time cancellation of muscle noise from the FECG signal is the expected result of this research. Successful application of this algorithm to this fetal electrocardiography problem would suggest other applications to problems in electrical and biomedical signal processing, since many problems of noise cancellation are difficult to solve due to the lack of a pure noise signal.
