Attitude Estimation & Control of a CubeSat Using Linear Quadratic Gaussian Approach

Abstract

Attitude estimation of satellites using Kalman Filters has been in practice for many years. The optimal attitude control in the presence of noise can be achieved by using the optimal controller and the optimal estimator, simultaneously. In this paper, the Linear Quadratic Regulator (LQR) has been implemented in conjunction with the Extended Kalman Filter (EKF) on a CubeSat model. Full quaternion-based model (dynamics & kinematics) of the CubeSat is employed for the design of LQR. Furthermore, an extended Kalman filter is designed using the reduced quaternion model. The filter is then implemented in the closed loop with the LQR, and the simulations are conducted. The data generation using the full quaternion model and the filter implementation using the reduced model, provide the benefit of computational ease all the while catering for any singularities in the model. The simulation results show adequate attitude control, estimation and noise filtration within a reasonable time and optimum control effort.