NANOSATELLITE ATTITUDE ESTIMATION IN SUN AND ECLIPSE PERIODS WITHOUT GYROSCOPES

The paper introduces attitude estimation methods for sun/eclipse periods without gyroscope measurements. Traditional attitude estimation approach with nonlinear measurements and non-traditional approach with linear measurements are investigated relying on sensor measurements from magnetometer, and sun sensor at the first stage. For the nontraditional approach, vectors are placed in Wahbas problem and the loss function is minimized by Singular Value Decomposition (SVD) method. The outputs of the SVD are used in the extended Kalman filter (EKF) in order to estimate the satellites attitude angles and angular rates. For the second stage in eclipse period, two additional methods are introduced as the prediction method, and traditional EKF based on only magnetometer measurements. For different phases, a switching algorithm is used for better attitude and rate estimation accuracy.