Suboptimal controller design for flexible launch vehicle based on genetic algorithm: Selection of the weighting matrices Q and R

In this study, a suboptimal controller for the pitch plane motion of a flexible launch vehicle is presented. Three sensors for measuring the motion, angular position sensor, angular rate sensor and accelerometer, are considered and assumed to be at given positions on a rocket described by a ninth-order state-space dynamical model, including three bending modes. The weighting matrices with respect to the suboptimal control method, which are used to synthesize a performance index, are selected by the genetic algorithm. Additionally, in order to improve the robust performance of the flexible launch vehicle, various uncertainty combinations are predetermined and considered in the process of the genetic search. The simulation results indicate that both the rigid body and flexible mode response are desirable, and the control performance including the robust performance is better than that of the suboptimal controller with manual selection of the weighting matrices.