Abstract
It has been reported that the accuracy of measuring the point of force application with force plate is affected by the position and magnitude of the force vector. The purpose of the present report is to evaluate the accuracy in determining the point of force application for investigating body sway, and to improve this accuracy by proposing a correction procedure. In the experiment with calibrated masses the measured points of force application were shifted parallel to X- and Y-axes in addition to being reduced comparing with the actual point. The error tended to be greater in the smaller magnitude of the masses. The root mean square of error (RMS error) at the smallest mass of 10kg was found to be 10mm X-coordinate and 6mm in Y-coordinate. The errors in the measured point of force application seem to be caused by 1) nonlinearity of load cells, 2) deformity of the top platform of the force plate and 3) differences in the gain and offset voltage among each load cell. The experimental results showed that the nonlinearity of the load cells was less than 2% even at the light loading of 2.8kg. A concave bending of the top platform due to force application makes a force act within the load cell rectangle. As a consequence, the measured value of force application tends to be greater than the actual point. Our results demonstrated the opposite, namely underestimation of the actual point. The greater deformity of force plate due to extreme force application like landing during running and jumping may become a significant factor. Thus, nonlinearity of the load cell and deformity of the top plate were ruled out as factors of the errors in investigation of body sway. A mathematical model of the force plate including load cells with different gain and offset voltage provided a satisfactory explanation of the errors in the measured point. Based on the observed error distribution and the model, a correction procedure independent of the magnitude of the force application was proposed. The correction algorithm improved the accuracy to the RMS error of less than 1mm at the calibrated mass of 10kg. It was confirmed that the correction procedure could provide more accurate and reliable information for investigating body sway.
