Estimation of Velocity and Acceleration using RBF Interpolation

Abstract

In this paper, an estimation method for velocity and acceleration from sampled displacement data using Radial Basis Function (RBF) interpolation is proposed. In some cases such as parameter identification of a robot using sampled signals of joint angle and motion control only with displacement sensor, accuracy of the estimation method for velocity and acceleratation affects the results of identification or control. In proposed estimation method, sampled displacement data are firstly interpolated with RBFs and estimated velocity and acceleration are derived from analytical differenciation of the interpolation. Off-line estimation that can be utilized in off-line parameter identification scheme is discussed. Considering on-line identification of System parameters of robot, quasi-instantaneous estimation method also developed, in which a least five latest measured data are required. Accuracy of the proposed method is evaluated in both the frequency domain and the time domain.