Haptic Data Compression by Curve Fitting

In order to reduce the distortion of offline haptic data compression, a compression method using curve fitting is proposed. Under the assumption that the haptic data are sampled from the original signals that can be represented by a curve defined by piecewise quadratic polynomials, a fitting curve of the haptic data is constructed by reversing the process of generating the haptic data, which is defined by piecewise quadratic parametric curves not only with first derivatives that agree at the points where they join, but also approximating the original signals with a quadratic polynomial precision. To avoid the storage of all coefficients of the fitting curve to further improve the data reduction rate, an iterative method for solving the coefficients of the fitting curve is derived so that only partial coefficients of the fitting curve are needed to be encoded and saved. Experimental results show that the proposed method can significantly reduce the compression distortion while maintaining a high data reduction rate.