Accurately estimate the real-time disturbance torque in the environment to improve the attitude control performance of the spacecraft. Based on flight data from a low-Earth-orbit satellite, an adaptive super-twisting sliding-mode observer (AST-SMO) is developed in this study, which integrates Kalman-filtered Euler-angle preprocessing for quasi-steady interval identification, adaptive band-pass filtering for angular-acceleration estimation, and a super-twisting observer with a state-dependent logarithmic gain law and dual leakage terms for chatter suppression. Compared with a fixed-gain observer, this one has a lower noise level and is more stable. The proposed observer improves the accuracy of disturbance-torque estimation and provides a practical basis for robust attitude-controller design.
