Attitude estimation for nanosatellites ssing Singular value decomposition and unscented kalman filter

2015-07-10
Söken, Halil Ersin
Hacızade, Cengiz
Çilden Güler, Demet
The singular value decomposition (SVD), one of the single frame attitude estimation methods, is used together with the unscented Kalman filter (UKF) for nanosatellite attitude estimation. The proposed interlaced algorithm estimates the attitude and gyro biases using three attitude sensors onboard, which are magnetometers, sun sensors and gyroscopes. The algorithm uses the SVD method at first to estimate the optimal attitude by minimizing the Wahba's loss function. In the second phase, the UKF takes this information as a measurement input for more accurate attitude estimates even in eclipse period in which the SVD method fails. The UKF runs by propagating only the kinematics, and the rotation angle error covariance matrix from the SVD is used as the measurement noise covariance matrix in the UKF. An attitude estimation accuracy that is within the desired limits for a nanosatellite mission has been achieved even when the satellite is in eclipse.
Citation Formats
H. E. Söken, C. Hacızade, and D. Çilden Güler, “Attitude estimation for nanosatellites ssing Singular value decomposition and unscented kalman filter,” presented at the 30th International Symposium on Space Technology and Science 2015, Kobe, Japonya, 2015, Accessed: 00, 2021. [Online]. Available: https://archive.ists.or.jp/upload_pdf/2015-f-37.pdf.