On-line attitude perturbation estimation in the earth-orbiting satellite

This paper presents three innovative techniques to estimate external perturbation torques acting on an Earth-orbiting satellite. The proposed techniques overcome the problem of derivative reconstruction from noisy measurement in the estimation procedure of the environmental perturbation torques. For this purpose, in the first technique, known as nonlinear disturbances observer (NDO) technique, derivative term is eliminated in the estimation equations by employing auxiliary variable vector and by selecting appropriate design parameters. In the second proposed approach, the robust exact differentiation via second-order sliding mode theory, known as Levant differentiator, is employed to decrease derivative estimation error in the process of estimation. To improve the performance of the Levant differentiator, the final innovative approach, known as sub-band adaptive robust differentiator (SARD) technique, is proposed in which the derivative term in the estimation process is reconstructed by using Levant differentiator and sub-band adaptive filter theory. In the SARD technique, the original noisy measurement signal is divided into sub-band signals and then, the Levant differentiator is applied to each sub-band signals to construct exact differentiator from noisy measurement. The novel aspect of paper is employing second-order sliding mode theory in the sub-bands frequency to improve the performance of derivative reconstruction in the presence of measurement noise. In this regard, a nonlinear model of the earth-orbiting satellite is simulated using specific navigation disturbances and the results verified the feasibility of the proposed strategies.