Robust Disturbance Observer-Based Control for Relative Attitude Tracking Using Sum-of-Squares Programming

摘要

This Note presents a computational approach for the synthesis ofnonlinear disturbance observers for relative spacecraft attitude tracking using sum-of-squares programming. Attitude tracking in thepresence of disturbance torques typically requires simplifying assumptionsfor disturbance observer design. These approximations are remediedby the method proposed in the Note by numerically computingthe observer gains based on the stability of linear time-varying systemsand using semidefinite programming. The stability of the disturbanceestimate error and the closed loop are studied using Lyapunov theory.Numerical simulations demonstrate the superior performance of theapproach as compared to baseline robust methods, such as slidingmode control, and with the observer design assuming approximatedynamics, even with fast time-varying disturbances.