Range limits of attitude determination accuracy
- Additional Document Info
- View All
This paper firstly provides the general mathematical expressions for the maximum, the mean, the variance and the standard deviation of the attitude error associated with an estimated attitude matrix A and the true attitude matrix T. Then, two mathematical tools are provided, which are able to estimate the maximum attitude error associated with a given attitude data set, to be used when n=2 and n>2 observed vectors are available, respectively. The maximum attitude error estimation is particularly useful because it represents an index of the robustness of the attitude estimation and, therefore, it can be used as a relative weight in statistical approaches (e.g. Kalman filters, optimal observers) and, in particular, as an indicator of the reaching of the "quasi parallel" condition of the observed vectors. The proposed maximum attitude error estimator is then validated by numerical examples highlighting its accuracy.
author list (cited authors)