Fit3dParametricCurv​e

Two nonlinear methods to fit multidimensional paths

Fits a nonlinear parametric vector path to simulated noisy 3d coordinate value observations with different normal distribution variance for each coordinate. To perform the fit, a single scalar function incorporating all three coordinates, with logical flags indicating the coordinate direction, is supplied to fitnlm. A maximum likelihood fit is also obtained using fmincon and gives consistent results. Consistency checks are performed by fitting one coordinate alone with fit and polyfit.

The techniques could be used for fitting similar related groups of observations with uncorrelated noise.