The likelihood is then given by $$ p(\vec{R}|\vec{x})\propto\exp\left[-\frac{1}{2}\left(\vec{R}-\mathcal{H}(\vec{x})\right)^{\top}\mathbf{\Sigma}_{obs}^{-1}\left(\vec{R}-\mathcal{H}(\vec{x})\right)\right] $$
Have you ever seen this function?
We can encode this as a matrix, if e.g. we stack $x$ over time in a vector...: $$ \begin{pmatrix} 1 &-1& 0 & 0 & \cdots\\ 0 &1 &-1 & 0 & \cdots \\ \vdots & \vdots & \vdots & \vdots & \cdots \\ \end{pmatrix}\vec{x} $$
It's a linear form!
eoldas_ng
¶eoldas_ng
tool