Robust Kalman Filtering and Its Applications.
Irwin Guttman, Daniel Peña · Defense Technical Information Center (DTIC) · 1984
This paper presents a robust Kalman filtering algorithm that is obtained assuming a scale contaminated normal distribution for the noise of the measurement equation. The mixture of normals obtained as a posterior distribution is approximated at each stage by a normal distribution with the same mean and variance. The resulting algorithm is simple, has a straightforward interpretation and seems to provide useful robust estimators in several statistical problems that are briefly reviewed. Originator-supplied keywords include: Robustness, and mixtures of normals.