- Generated by
1.18.0
|
EmrysFilter 0.1.0
State estimation utilities in C++23 using Eigen
|
Linearized method: \( \bar y = f(\mu) \) and \( \Sigma_{yy} = \sum_i J_i \Sigma_i J_i^T \), with the Jacobians \( J_i \) of f at the input means. More...
#include <EmrysFilter/Models/Methods.h>
Linearized method: \( \bar y = f(\mu) \) and \( \Sigma_{yy} = \sum_i J_i \Sigma_i J_i^T \), with the Jacobians \( J_i \) of f at the input means.
Exact for an affine f.
An f that returns its Jacobian with its value (withJacobian()) is called once, at the means, and that Jacobian used. Any other f is differentiated by forward-mode automatic differentiation, with Eigen's AutoDiffScalar, and must be generic in its scalar type; one that returns plain scalars when given dual numbers fails with a static_assert that says so. There is no numerical differentiation by fixed steps: no one step size suits every input, and the covariance, which should set it, is what the unscented methods already spread their points by (decisions.md [returned-jacobians]).
1.18.0