Fuzzy-scored genetically-designed controller for the VLS-1 launcher

A fuzzy system intertwined with an optimization technique based on genetic algorithm for designing a VLS-1 Launcher attitude control system is proposed in this paper. A linear-quadratic optimization yields gain-scheduled controllers, where Q and R weighting matrices are scored by means of a fuzzy system, instead of directly being chosen by a human designer. As a consequence, not only the optimization technique is extended to the entire launcher simulated trajectory, but also the specifications can be dealt individually. The linguistic variables of the fuzzy system are defined according to performance (rise time and overshoot) and robustness (gain and phase margins) issues, regarding gain vector smoothness, as well. The control system obtained with fuzzy design presents many favorable results according a linear analysis and nonlinear digital simulations.