56
Hembree & Zirnmer
where Gt = P; H’(R< + H P; H’)'1 and is known as the Kalman gain. Notę that this gain is a blending parameter that relatively weights the prior estimate and the current measurement.
The algorithm for performing the recursive estimation of the Kalman filter is shown in FigurÄ™ 2. We start with prior estimates for the State and its error covariance. From this we calculate the gain. Next we compute the current estimate of the State using the calculated gain and the current measurement. Next we compute the error covariance for the current estimate. Finally we propagate the estimate and its covariance to use as priors for the next step.
FigurÄ™ 2. Kalman Filter Algorithm.