Signal averaging in stress test and Holter systems

Abstract

A signal averaging technique is presented as a high quality alternative for filtering and cleaning ECG signals. A basic mathematical approach is first presented, with emphasis on the bandwidth reduction due to imperfect synchronism, when the QRS complex and R wave peak are used for synchronizing and alignment. A very reliable method is presented, especially designed for noisy ECG signals; in this method several thresholds are generated, based on the history of the successive cycles; thresholds for the amplitude of high pass filtered ECG signals and their first derivatives are generated in an adaptive and stable way. Finally, results are shown for several typical cases, including static averaging and dynamic real-time averaging.<>