On May 22, 2018, the Gravity Recovery and Climate Experiment Follow On (GRACE-FO) mission was launched with the goal to map the spatiotemporal variations in the Earth's gravity field and to extend the 15-year monthly mass change observations of its predecessor, the GRACE mission.
...
On May 22, 2018, the Gravity Recovery and Climate Experiment Follow On (GRACE-FO) mission was launched with the goal to map the spatiotemporal variations in the Earth's gravity field and to extend the 15-year monthly mass change observations of its predecessor, the GRACE mission. Similarly to GRACE, the measurement principle of GRACE-FO is based on three different key elements, namely inter-satellite ranging, precise orbit determination and accelerometry. The accurate estimation of the satellites’ attitude has an influence on all three of them. Therefore, any unmodeled errors in the attitude dataset product can propagate to the gravity field solutions and degrade the results. The objectives of this thesis are twofold. Firstly, to analyse the in-flight performance of the GRACE-FO star cameras, fiber-optic gyroscopes, accelerometers and steering mirrors. Secondly, to propose a method that accounts for the instruments' noise and errors and fuses the data, giving an improved attitude solution.
The noise and error characteristics of each instrument are determined by examining their measurements in the time and frequency domain, as well as investigating their differences in geographic plots. These analyses are performed for May 2020, when all instruments could provide nominal data. Motivated by the improved attitude quaternions and gravity field results that were obtained from the latest GOCE gradiometer data calibration process, the proposed on-ground attitude reconstruction for GRACE-FO is composed of three elements. The first one is the optimal combination of star camera and steering mirror quaternions by minimizing the weighted residual sum of squares of the elements of the noise quaternions. Within this combination, a set of constant parameters are also estimated that describe the relative alignment of these sensors. The second element is the reconstruction of the satellite angular rates in the frequency domain by applying respectively a highpass and a lowpass filter to the IMU and to the combined star camera and steering mirror derived angular rates. Lastly, the third element is the attitude reconstruction, for which attitude quaternions, resulting from the smooth reconstructed angular rates, are fitted to the optimally combined quaternions by means of a generalised least-squares adjustment.
The proposed attitude data fusion method produces an improved attitude solution that incorporates more accurately the noise and error characteristics of the star camera, the steering mirror and the IMU measurements. At the level of quaternions, it performs better than the official method, which is based on Kalman filtering, with noticeable improvements at frequencies above 10 mHz. However, based on a comparison of the corresponding derived antenna offset correction for range rate, very minor improvements are expected at the level of the gravity field. This is due to the K/Ka-band ranging system noise being the dominant source at the higher frequencies. The findings of this thesis work are valuable for the design of future gravity missions such as the Next Generation Gravity Mission proposed by ESA, for which a redundant accelerometer design is considered. Given the estimated noise characteristics of the above instruments and the proposed angular rate reconstruction method, the most favorable placement of the accelerometers is found to be in the along-track direction. If an accelerometer fails in this configuration, the noise in the required centrifugal and Euler acceleration corrections will be less than that of the laser ranging system.
