DETERMINATION OF OPTIMAL FILTER PARAMETERS FOR FILTERING KINEMATIC WALKING DATA USING BUTTERWORTH LOW PASS FILTER

Abstract

INTRODUCTION: Noise in biomechanical displacement data are often reduced by digital filter routines. In literature the cross-validated quintic spline from Woltring (1986) and the Butterworth low pass filter are commonly used. Challis (1999) recently suggested a method for determining the optimal cut-off frequency of Butterworth low pass filters. This method is based on a combination of residual analysis and autocorrelation. The optimization of the filter order, however, is not discussed by Challis. Thus the purpose of this study was to determine optimal filter parameters (frequency and order) of a Butterworth low-pass filter using the approach of Challis applied to a kinematic data set in walking.