Force prediction using singular-value decomposition

Abstract

Results of a preliminary investigation into a method which can be used to improve the results of an experimental force prediction analysis is presented. The method uses a singular value decomposition technique to reduce measurement noise present in the measured input data matrix. The fundamental theory for the method along with an experimental example is shown. The method is shown to provide encouraging force prediction improvements when the input data matrix contains high levels of measurement noise.