Class StateCovarianceBlender

  • All Implemented Interfaces:

    public class StateCovarianceBlender
    extends AbstractStateCovarianceInterpolator
    State covariance blender.

    Its purpose is to interpolate state covariance between tabulated state covariances by using the concept of blending, exposed in : "Efficient Covariance Interpolation using Blending of Approximate State Error Transitions" by Sergei Tanygin.

    It propagates tabulated values to the interpolation date assuming a standard keplerian model and then blend each propagated covariances using a smoothstep function.

    It gives accurate results as explained here. In the very poorly tracked test case evolving in a highly dynamical environment mentioned in the linked thread, the user can expect at worst errors of less than 0.25% in position sigmas and less than 0.4% in velocity sigmas with steps of 40mn between tabulated values.

    Vincent Cucchietti
    See Also:
    SmoothStepFactory, SmoothStepFactory.SmoothStepFunction