SPECTRAL DECOMPOSITION OF VERTICAL GROUND REACTION FORCE CURVES

Abstract

INTRODUCTION - The vertical ground re- action force component typically has two peaks. The first peak (impact peak) is caused by the impact between the foot and the ground while the lower frequency second peak (active peak) is caused by the vertical braking of the body followed by vertical push-off. Bobbert et al. (1991) pro- posed a method of decomposing the VGRF into the contribution of the support leg and the rest of the body by double differentiation of segment center of mass position data. This decomposition allowed them to determine the magnitude of the impact peak independent of the rest of the curve. The purpose of this study was to investigate a method of decomposing the VGRF curve that does not require differentiation of position data. METHODOLOGY - Five male recreational runners completed 5 trials in 3 different run- ning shoes that differed only in the density of the midsole material. VGRF data were recorded from a force platform at 1000 Hz using an analogue to digital converter. A Fourier transform was performed on each trial An inverse transform was then per- formed twice - once using the frequencies below 3 Hz and once using the frequencies above 3 Hz. Figure 1 illustrates a typical decomposition of the force curve. Peak (PK) and time to peak (TPK) values were recorded for each curve of the decomposed VGRF curve. RESULTS AND DISCUSSION - The magnitudes and times of the impact and active peaks for the decomposed VGRF are presented in Table 1. The impact peak for the soft shoe had a greater magnitude than the medium or hard shoes. This may indi- cate that this shoe midsole experienced maximum compression or that the subjects perceived the soft midsole and adjusted their kinematics. There were essentially no differences in the active PK between the soft, medium and hard midsoles. Table 1. Mean values for decomposed VGRF variables. Midsole Soft Medium Hard Impact PK 240.1 212.8 238.7 Impact TPK 28.3 26.1 26.6 Active PK 1241.2 1233.4 1244.7 Active TPK 107.9 108.7 104.0 Since the impact peak component is of a higher frequency than the active compo- nent, the decomposed VGRF curves are comparable lo the curves of Bobbert et el. (1991). These results indicate that this pro-cedure can be used to separate the true impact characteristics of the VGRF from the remainder of the curve. REFERENCES Bobbert, M.F. et al. (1991). Calculation of vertical ground reaction force estimates during running from positional data. J Biomech, Vol 24: 12 pp. 1095-1 105.