NOVEL MULTIPLICATIVE UNSCENTED KALMAN FILTER FOR ATTITUDE ESTIMATION

A novel spacecraft attitude estimation algorithm is presented. The new algorithm utilizes unit vector measurements and is based on the unscented Kalman filter (UKF). The UKF, like the extended Kalman filter, employs a linear update in which an additive residual is formed. The residual is given by the difference between the measurement and its mean. This work utilizes a multiplicative residual in which the measurement and the mean are multiplied together using the vector cross product. Because of the nature of the problem, a multiplicative residual combined with a multiplicative update is a more natural solution.

NOVEL MULTIPLICATIVE UNSCENTED KALMAN FILTER FOR ATTITUDE ESTIMATION Conference Paper