Abstract
One method to calculate sensor accuracy necessary for Kalman filter is presented. It is an inverse problem of the Riccati equation in Kalman filter, and observation sensor or input sensor accuracy is searched numerically to achieve the specified estimation error.