On Computational Efficiency of Linear Filtering Algorithms.
Abstract
Although considerable attention has been given to various aspect of the filtering problem, the question of computer requirements has not been analyzed in detail. The report presents an attempt at a systematic analysis of this problem. Formulae for the arithmetic operation count and storage requirements are derived for several popular filtering algorithms within the framework of a general estimation problem. Parameters included in the study are the dimensions of the state vector, the frequency and number of observations, and the frequency of state vector updates. The results for each algorithm are discussed from a practical viewpoint and comparisons are made for certain interesting cases. Several common misconceptions are exposed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 15, 1970
- Accession Number
- AD0718410
Entities
People
- A. B. Bierman
- I. A. Gura
Organizations
- The Aerospace Corporation