Abstract

A large spinning spacecraft in Earth orbit includes a rigidly mounted telescope, parallel to the spacecraft's intended spin-axis. The three principal moments-of-inertia are unequal. Electric thrustors are used to overcome the gravity-gradient torque components. Uncertainty of the spacecraft's inertia tensor, including misalignment of the telescope axis from the actual principal direction, as well as magnetic, solar pressure, and aerodynamic forces, produce disturbing torques which are constant or vary sinusoidally during a spin-rotation. The constant torques and the coefficients of the sinusoidal torques are modeled as first-order Markov processes and consequently increase the dimension of the dynamical state. Measurements of the spacecraft's orientation are assumed to be of 2 types: a) the deviation, measured continuously in body axes, of the telescope axis from its desired direction; b) the deviation of the telescope axis towards or away from certain bright stars, measured as the star in question crosses a slit on the outside of the spacecraft. The estimation of the orientation from measurements a) and b) involves not only linear dynamical equations with periodic coefficients but also a mixture of a continuous and a discrete periodic, information pattern. The usual estimation gains settle here into a periodic pattern. Quadratic synthesis of a control law produces constant and periodically varying feedback gains. The constant control gains are obtained by eigenvector decomposition, and the periodic control gains by solving a linear matrix differential equation with periodic forcing term. The method of spectral factorization is used to obtain the estimator gains.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call