Using Cross-Entropy to Assess Accuracy of Position Estimators
- Paper number
IAC-15,B2,IP,10,x27821
- Author
Mrs. JEANNETTE NOUNAGNON, Virginia Tech, United States
- Year
2015
- Abstract
In regard to assessing the accuracy of position estimators, although limitations in using Mean Squared Error (MSE) are known, comparing MSEs has, for a long time, been the most accepted approach. In this paper, we introduce the use of cross-entropy as a very robust alternative. We show the limitations of using MSEs and validate via simulation results that using a comparison of cross-entropies to identify the most accurate estimator is more robust than a comparison of MSEs.
- Abstract document
- Manuscript document
(absent)