ABSTRACT
The algorithm for optimal fixed point data smoothing for continuous linear systems is developed by solving the appropriate Wiener-Hopf matrix integral equation. Two formulations are given for generating the optimal smoothing filter gain matrix. It is found that one of the formulations is preferable for computations since it avoids inversion of a matrix that may be ill-conditioned, a feature inherent in the other formulation. Three equivalent expressions are derived for the determination of the smoothing error covariance matrix. The results substantiate and extend those obtained previously for the same problem via another method.
Notes
† Communicated by the Author.