Retrospective Cost Adaptive Control of Spacecraft Attitude Using Magnetic Actuators

We apply retrospective cost adaptive control (RCAC) to spacecraft attitude control with magnetic actuators. We compare two approaches to address the rank deficiency and time-varying nature of the input matrix. The first approach utilizes an average of the magnetic field based on a-priori knowledge, whereas the second approach uses three multi-input, single-output controllers. RCAC uses no information about the spacecraft inertia, and model information is limited to the input-output relation given by the first Markov parameter, which is computed from an inertia-free linearization of Euler’s and Poisson’s equations. We examine two problems for each of the controllers. For both problems, the spacecraft has an arbitrary initial angular rate and initial attitude. The objective for the first problem is to bring the spacecraft to rest at a specified attitude, while the second problem seeks to bring the spacecraft to spin about an inertially pointed body axis.