Fitting Biomechanical Data with Cubic Spline Functions

Abstract

A systematic comparison of orthogonal polynomial and cubic spline function data modeling techniques was conducted for empirical biomechanical data. Kinetic data derived from mathematical modeling and cinematographic techniques, using either polynomial or spline curve fitting and differentiation, were compared to simultaneously recorded force platform data. Results indicated a superiority of the spline functions for modeling the empirical data. Mean percent difference between vertical ground reaction forces recorded by the force platform and the film derived data was less than 5% when spline curve fitting techniques were used. Orthogonal polynomial curve fitting techniques produced a mean percent difference greater than 10% for the same comparison.