---

This paper presents a new method to design optimal thrusters’ configuration for geostationary satellite in order to reduce the fuel consumption and increase the control accuracy. The thrusters configuration generally contains information about thrusters fixed on the satellite body structure, including their location, orientation. One important factor playing a key role in thrusters’ configuration design is satellite force-torque analysis. The proposed configuration, however, should lead to fulfill specified attitude maneuver when the set of force and torque produced by satellite thruster system is adequate. For this purpose, two optimization methods using genetic algorithm (GA) and differential evolution (DE) has been applied to determine the optimal thrusters configuration on the communication satellite body. The cost function employed to minimize both the fuel consumption and error generated by thrusters installation and uncertainties. Moreover, this work allows applying some different constraints in the proposed formulation including minimization of the thruster plume impingement effect on the satellite outer structure and on the solar arrays and the second one is the satellite dimension and geometry. Simulation results show that DE outperforms GA in terms of accuracy and CPU time. Effectiveness of differential evolution algorithm is illustrated in the paper when compared with GA results.

---

In this paper, LQG/LTR controller is designed for attitude control of the geostationary satellite at nominal mode. Usage actuator in this paper is the reaction wheel and control torque is determined by the LQR regulator. Usage sensors in this article are sun and earth sensors and EKF are used for estimation of noisy states. LQR controller signal has good performance, if all system's states are considered in system output feedback. But this method is ideal and does not include model noise and sensors noise. Therefore, LQG and LQG/LTR controllers are designed based on the estimated states, and are compared with LQR controller. Controllers gain coefficients are obtained based on linearization about working point. It caused to robustness and similarity of LQG and LQG/LTR response. The results show that control overshoot of LQR is greater than the others.

---

In this paper, a LQG/LTR controller is proposed for attitude control a geostationary satellite at nominal phase. Basically, proposed methodology includes three parts: LQR regulator, EKF, and loop transfer recovery. Controller design is based on the linearized equations of the spacecraft dynamics using reduced quaternion model. Reduced quaternion model solve uncontrollable problem in some subspaces in the linearized state space quaternion model using all four components of quaternion. Spacecraft actuators are reaction wheels and attitude determination sensors are sun and earth sensors. LQR controller is ideal and it doesn’t account for the model uncertainty and sensor noise and it uses the feedback of the full states. To consider the model uncertainty and sensor noise, we have designed EKF which is used by LQG and LQG/LTR controllers. Controller gain coefficients are obtained using a reduced quaternion model, and based on linearization around the equilibrium point and the natural frequency of the closed loop system. To increase the robustness of the design with respect to solar radiation disturbance, singular values of LQG are approximated to Kalman filter, in LTR section. The results demonstrate that LQG/LTR performance is better than LQG’s and LQG/LTR has a good robust stability margin with respect to disturbances.