An Introduction to the Kalman Filter
An Introduction to the Kalman Filter
Optimized Bayesian Dynamic Advising: Theory and Algorithms (Advanced Information and Knowledge Processing)
Particle filtering with factorized likelihoods for tracking facial features
FGR' 04 Proceedings of the Sixth IEEE international conference on Automatic face and gesture recognition
Hi-index | 0.00 |
The paper deals with a factorized version of Kalman filter. Via factorization of a state-space model such the filter provides the state estimates of individual state vector entries in the factorized form and allows to update them entry-wise. The paper continues a series of research in the field of the factorized filtering and proposes the novel modified algorithm, including the simultaneous entry-wise organized fulfillment of data and time updating steps. A motivation of the research is a preparation of the universal algorithm for the joint filtering of variables of a mixed (continuous and discrete-valued) type. The illustrative example and comparison of computational complexity with other versions of Kalman filtering are presented.