Robust Optimization of Satellite Attitude Control System with On-off Thruster Under Uncertainty

In this paper, a modified PI-D controller with pulse width pulse frequency (PWPF) modulator with on-off thruster for a rigid satellite is tuned under uncertainties in a quas-inormalized form. A modified proportional-integral-derivative (PID) based on observer method is used as a controller. Uncertainty is considered on thruster model, thrust level, and external disturbance parameters. Thruster is modeled with a constant delay followed by a second-order binomial transfer function. The modulator update frequency, the input to the on-off thruster, is limited to 50 Hz. Controller gains are tuned and optimized using a multi-objective genetic algorithm. The simulation results of the deterministic optimization show that the rise time is reduced and the overshoot is removed. Moreover, robust optimization is applied to this problem using a combination of deterministic optimization and Monte Carlo simulation. Next, the statistical properties of the results, such as expected value and variance are evaluated. These properties are used as the objective functions of the optimization problem, and the results are compared to each other. The variation of objective function in robust optimization is reduced, which means that the robustness under uncertainties is enhanced. Using quasi-normalized equations and merging the moment of inertia, thrust level and filter gain with other parameters, the results are more general and useful for other spacecraft with different characteristics.