Automatic mass balancing system for a dynamic CubeSat attitude simulator: development and experimental validation

This paper describes the automatic balancing and inertia identification system for three degrees of freedom CubeSat attitude simulator testbed. For a reliable verification of the attitude determination and control subsystem, the on-orbit environment shall be simulated within the testbed, minimizing the external disturbances acting on the satellite mock-up. The gravity torque is expected to be the largest among the disturbances, and an automatic balancing procedure can largely reduce the time necessary for tuning the platform and minimize the residual torque. The automatic balancing system adopted in this work employs three sliding masses independently actuated by three electric motors using a two-step procedure. In the first step, a feedback control is employed for a plane balancing. The inertia parameters and the remaining offset component are then estimated by collecting free oscillating platform data. This two-step procedure is iterated towards increasingly finer balancing until no further improvement is obtained. For the planar balancing, a control law based on linearized equations and a newly developed nonlinear feedback law is implemented and compared, showing the superior performance of the latter. The unbalance offset vector component along the local vertical and inertia tensor are estimated by a constrained batch least squares filter. Experimental results show the effectiveness of the implemented approach, which leads to a residual disturbance torque acting on the balanced platform smaller than 5 × 10–5 Nm.