Methods for estimating spacecraft inertia Tensor

In 2011, the National Space Organization (NSPO) of the National Applied Research Laboratories (NARL) in Taiwan, Republic of China (ROC), initiated a latest space program, called the FORMOSAT-7 (FS-7), whose mission consists of a new constellation of 12 satellites (to be built by Surrey Satellite Technology LTD) plus one additional NSPOBuilt satellite for an operational mission to replace the previous FORMOSAT-3 (Cosmic) mission. The NSPO-built satellite is a 3-axis attitude stabilized spacecraft with a single solar wing attached to the spacecraft main body. The single solar wing has a 2 DOF (degree-offreedom) motion: track motion and trim motion to track the Sun at various Sun beta angles during normal operations. Due to this highly non-symmetric body, the spacecraft 3x3 inertia tensor will contain large non-zero cross-coupling terms. In addition, the values of the inertia tensor will vary as the solar wing´s track angle and trim angle change. The issues of large non-zero cross-coupling terms and the timevarying inertia tensor post a challenging bus Attitude and Orbit Control System (AOCS) design. We propose a MultipleInput Multiple-Output (MIMO) controller for the FS-7 bus attitude control using de-coupling control technique to deal with the large non-zero cross-coupling terms and real-time spacecraft inertia tensor estimates to cope with time-varying inertia tensor. This paper describes methods for spacecraft inertia tensor estimates for the purpose of supporting the proposed MIMO control design architecture. The derivations of this spacecraft inertia tensor estimator are the extension of the previous work, where a gyro less attitude and rate determination algorithm was derived using the computed gyro data (numerical gyro data). With the obtained linearized attitude error equations and the linearized rate error equations/spacecraft inertia tensor error equations, a 12-states Extended Kalman Filter (EKF) providing spacecraft attitude estimates, rate estimates, an- inertia tensor estimates is then implemented. A Matlab-based covariance analysis simulation model was developed to assess the filter performance. Their performance sensitivities to various spacecraft maneuvers will be presented in the paper. The algorithm will be tested in a 6-DOF nonlinear, high-fidelity simulation model to be developed in the future for the FS-7 program to assess its ultimate performance.