Adaptive Spacecraft Attitude Control with Actuator Saturation

P RACTICAL spacecraft attitude control systems must operate in the presence of disturbances, modeling errors, and actuator limitations. These issues have been the subject of much research interest. Adaptive control, where the unknown system parameters are estimated adaptively, is one of the proposed approaches for dealing with modeling uncertainty (see, for example, [1–4]). Both [1,2] deal with the attitude tracking problem, but they do not treat disturbances or actuator saturation. Reference [3] also deals with the tracking problem; it includes actuator saturation but not disturbances. Reference [4] includes bounded disturbances, but it does not treat actuator saturation, and it only deals with the attitude regulator problem. Recently, new control laws have been obtained that treat both disturbances and actuator limitations simultaneously [5–8]. References [5,6] deal with the attitude regulation problem only. References [7,8] treat the attitude tracking problem, and both present globally convergent control laws, given bounds on the spacecraft inertia matrix and the disturbances. The advantage of these approaches is that the form of the disturbance need not be known, only the bound. On the other hand, these approaches have no ability to learn the system model, which could be a useful feature if the attitude motion is to be optimized. This Note shows that, when an adaptive attitude control law based on the form given in [2] is appropriately designed, any linearly parameterizable disturbances can be accommodated; the closed-loop system is stable, with asymptotic tracking in the presence of actuator saturation. The unknown system parameters are learned adaptively.