unscented_transform_mean_covar

ssapy.utils.unscented_transform_mean_covar(f, x, C, scale=1)[source][source]

Compute mean and covariance matrix using unscented transform given a transformation f, a point x, and a covariance C.

This uses the sigma point convention from sigma_points. It assumes that f(sigma_points)[i] is f evaluated at the ith sigma point. If f does not obey this convention, this function will produce undefined results.

Parameters:
  • f (function) – the function to evaluate at the sigma points.

  • x (array_like (n)) – the central value to evaluate the function around

  • C (array_like (n, n)) – the covariance matrix corresponding to x

  • scale (float) – return scale-sigma points rather than n-sigma points. e.g., for 5 sigma, set scale = 5.