In-orbit calibration of drag-free satellites

In the near future a series of scientific satellite missions on fundamental physics as well as geodesy will be launched. A significant number of them will utilise a system for disturbance reduction, a so called Drag-Free Control (DFC) system. This new key technology allows to reduce the residual accelerations on experiments significantly. Its concept involves a free-flying proof mass inside the satellite which is shielded from non-conservative disturbances by the satellite. The Drag-Free Control system forces the satellite to follow the proof mass in order to generate a low disturbance free-fall environment. In order to provide a very low disturbance environment (for some missions < 10(-14) m/s(2)) the Drag-Free Control system has to be optimised. Its quality is limited by the level of knowledge about the system to be controlled. The model and its parameters used for the control system design determine the control accuracy. If the model is inaccurate or uncertain the control quality will be poor. So there is a need to calibrate the system. In this sense calibration means the identification of the system for optimal state estimation and optimal control. The complex dynamics of the proof mass-satellite system makes it impossible to carry out a proper system identification for a Drag-Free Control system on ground. The identification can only be carried out in orbit. But special algorithms have to be prepared in order to improve the model and to derive the parameters more accurately. This paper will address the general issue of in-orbit identification of drag-free satellites. It will give an overview on expected uncertainties and their effect on the dynamics of the system. Different approaches for the identification will be discussed. A first approach is presented. (c) 2005 COSPAR. Published by Elsevier Ltd. All rights reserved.