Citation Context

...e “noise” in thescoordinate data, mainly due to the digitizing process,shas been calculated to have an rms error of 2 mm. Priorsto link segment energy analyses the coordinates weresdigitally filtered =-=(23)-=- using a fourth-order, zero-lag, low-spass Butterworth filter cutting off at 5 Hz. Validationsfor the filtering and finite difference calculation ofsvelocities and accelerations is supported by the st...

Citation Context

...n higher frequencies, consequently, motion filtering uses low pass filter. (a) Low pass. (b) High pass. (c) Band stop. (d) Band pass. Fig. 12: Types of square filters. The low pass Butterworth filter =-=[33]-=- is applied to a forward time series x1, x2, …, xn obtaining a filtered time sequence y1, y2, …, yn . The filter is then reapplied to the reverse RITA • Volume XII • Número 2 • 2005 19sOptical Motion ...

Citation Context

...MSE Robustness to Noise in Motion Capture Most studies of aimed movements have employed measurement instruments in which the level of noise is low. Noise is inevitable in motion capture data, however =-=[20]-=-. Ideally, the metric would tolerate a level of noise that does not obscure controlled aspects of movement. To understand the effect of noise, we added white noise with zero mean and variance to each ...

Citation Context

...re body. T. Yanai / Journal of Biomechanics 37 (2004) 605–612 607Performance Technologies, Denver, CO, USA) for one stroke cycle. Twenty-one body landmarks were digitized in each field to represent the end points of the head, torso, upper arms, forearms, hands, thighs, shanks and feet. The resulting sets of two-dimensional coordinate data were used to determine the corresponding threedimensional coordinates, and expressed with respect to a global reference system embedded to the pool deck at the water surface. The three-dimensional coordinates were smoothed using a Butterworth digital filter (Winter et al., 1974) with various cut-off frequencies (2–4Hz) selected to retain at least 95% of the power of the raw signal. The buoyancy torque was determined with the method described by Yanai (2001a). In short, this method uses the positions and orientations of all body segments determined by three-dimensional videography and the dimensions of each body segment estimated on the basis of body segment parameters reported in the literature (density and centroid position—Drillis and Contini, 1966; mass—Clauser et al., 1969; position of the center of mass—Hinrichs, 1990). The volume of the torso was estimated as t...

Citation Context

...ed at 200 Hz from two adjacent multi-component force platforms (9286A, 9281C; Kistler Instrumente AG, Winterthur, Switzerland). At the outset of each trial the participants were instructed to step off the platform for zeroing to reduce sensor drift during prolonged stationary periods. Motion and analog ground reaction force data were recorded using Vicon Workstation and then exported to and analyzed with Visual3D, version 3.79. The digitized marker coordinates were filtered with a fourth-order Butterworth low-pass digital filter (cf. Pezzack, Norman, & Winter, 1977; Robertson & Dowling, 2003; Winter, Sidwall, & Hobson, 1974) with 4 Hz cutoff frequencies, due to the low frequency content of the data. The analog force platform signals were filtered with a fourth-order Butterworth lowpass digital filter with a cut-off frequency of 10 Hz. The joint and segment kinematics and force platform signals were incorporated into a standard link segment model (Bresler & Frankel, 1950) to compute joint angular velocities (rad/s) and body mass normalized, moments of force (N m/kg) and moment powers (W/kg) at the ankles, knees, and hips of both lower extremities. Joint angular velocities, moments, and power were projected onto l...

Citation Context

...ut data and a Fourier series fit of the motion trajectories. Visual inspection of a plot of variance versus number of the harmonic used in the fit was employed, but only an "arbitrary criterion" could be used to finally determine the highest harmonic retained. The numerical techniques for data smoothing noted in the foregoing are all operations that can be implemented as digital filters. Recent advances in digital signal processing techniques have made it possible to implement algorithms which are fast, accurate and well characterized with respect to overall frequency response. Winter, et al. [7] were among the first to employ the techniques of modern digital signal processing for the reduction of noise in anatomical point trajectory data obtained during walking. The displacement data were obtained from a computer-video system. By examining the frequency spectrum of anatomical point trajectories it was determined in [7] that most of the power in the signal was below the seventh harmonic. The figure varied depending upon the point being considered. To determine if the data above a given harmonic were noise, a spectral analysis of position data for the markers on the heel and ball of th...