AUTHOR
Blomqvist, Johan

DEPARTMENT
Space Science, Kiruna

SUMMARY
The attitude sensors for small spacecraft are often selected from commercial components to reduce costs. Uncertainties in these sensors hardware contribute significantly to the accuracy of the attitude solution. The idea that Linear Covariance techniques can be used to predict the accuracy of attitude determination systems and assist in their design is investigated. By using the sensor’s parameter estimated accuracy in a model, one could calculate the total standard deviation of the attitude determination by propagate errors through the estimation and use simple Root-Sum-Square. The linearization of the estimation process as a function of error sources can be done automatically using software linearization tools. Generalized Matrix Laboratory (MATLAB) M-functions using this technique are written in order to provide a tool for estimating the attitude determination accuracy of a small spacecraft. This tool were applied to a satellite dynamics truth model developed at Satellite Dynamics Laboratory, UTAH, USA, in order to quantify the effects of sensor uncertainties on this particular spacecraft’s attitude determination accuracy. Many error sources, including sensor mounting, bias drift, sensor noise and thermal fluctuations were examined. As expected, primary contributors to the error were identified. Design actions such as other processing methods, sensor calibration and alignment methods to meet the required system specifications were suggested based on the analysis.

ISSN 1653-0187 / ISRN LTU-PB-EX--10/002--SE / NR 2010:002