ESA GNC Conference Papers Repository

Very high accuracy attitude determination for LOS steering
M. Giurato, F. Haydar, M. Lovera, G. Sechi
Presented at:
Salzburg 2017
Full paper:

A challenge for instrument control systems is to meet increasingly stringent attitude pointing and/or knowledge requirements imposed by emerging advanced science and Earth observation payloads. One possible approach for improving attitude pointing performance is to provide an active compensation of the disturbances, using sensors having bandwidths beyond those of the standard 3-axis suite of a gyroscope and star-tracker and mounted inside or close to the payload itself. In this study, an approach to line-of-sight sensing based on using a high-bandwidth Angular Rate Sensor (ARS), together with attitude sensors like a star tracker and a gyroscope is considered. In this work, an ARS is defined as a rate sensor the useful bandwidth of which extends up to 500 Hz and higher, i.e., well beyond that of a conventional and available gyroscope, and without the additional accommodation issues provided by accelerometers. The ARS sensor based on a magneto-hydrodynamic concept, developed by Applied Technology Associates (ATA) has been considered. A peculiarity of the ARS is that the phase delay and the magnitude are not constant in the frequency range where the sensor is used, providing distortion on the measured angular rate. Following a review of the existing literature, see references [1-2], parametric models for the dynamics of the ARS have been identified from manufacturer data, taking also into account model uncertainty due to temperature dependence and frequency-domain fitting errors. Based on such models, a fully integrated solution to the three-axis attitude and rate determination problem has been developed, both for on-line and off-line filtering. In particular, the developed multiplicative-extended filter and smoother can be tuned using either a Kalman filtering or an approach, based on dedicated tuning guidelines which have been developed for both filtering/smoothing frameworks. The performance of the developed algorithms, to be reported in the complete paper, has been assessed in a simulation study, based on test cases representative of desired filter bandwidth in the order of 1 kHz. The results of the simulation study confirm findings from the relevant literature about the high accuracy and wide bandwidth of the achievable attitude and angular rate estimates. With respect to the existing literature, however, the study takes also into account model uncertainty associated with the ARS frequency response function. In addition, the time and frequency-domain analysis of the results has led to an improved insight in the role of multirate filter/smoother implementation in determining the attitude and rate estimation performance.