Accession Number : AD1027127

Title :   Cubature/ Unscented/ Sigma Point Kalman Filtering with Angular Measurement Models

Descriptive Note : Conference Paper

Corporate Author : Naval Research Laboratory Washington United States

Personal Author(s) : Crouse,David F

Full Text :

Report Date : 06 Jul 2015

Pagination or Media Count : 8

Abstract : Filtering algorithms that use different forms of numerical integration to handle measurement and process non-linearites, such as the cubature Kalman filter, can perform extremely poorly in many applications involving angular measurements. We demonstrate how such filters can be modified to take into account the circular nature of the angular measurements, dramatically improving performance. Unlike common alternate techniques, the cubature methods can be easily used with angular measurements arising from ray-traceable propagation models.

Descriptors :   kalman filters , gaussian distributions , algorithms , weighting functions , probability distributions , normal distribution , statistical algorithms , covariance , measurement , polynomials , estimators

Distribution Statement : APPROVED FOR PUBLIC RELEASE