A variable step size LMS algorithm for the suppression of the CPR artefact from a VF signal

Artefacts created by thoracic compressions during cardiopulmonary resuscitation (CPR) prevent the proper classification of the cardiac rhythm by an automatic external defibrillator (AED), making a pause in CPR necessary for a correct rhythm analysis. Previously proposed adaptive filtering methods have produced satisfactory CPR cancellation results but involve complex out of hospital intervention scenarios. A new adaptive method requiring minimal modifications of the basic operation of an AED is proposed for the suppression of the CPR artefact from a ventricular fibrillation (VF) rhythm. A model of the CPR artefact, based on the instants of the thoracic compressions, is used as the reference signal of a variable step size least mean squares (LMS) algorithm. Emphasis is put on building a reference signal highly correlated to the CPR artefact and on the fine tuning of the LMS algorithm. Satisfactory results have been obtained for the average increase in signal to noise ratio (SNR) and the sensitivity of a commercially available AED