Robust Control of a Reusable Launch Vehicle in Reentry Phase Using Model Following Neuro-Adaptive Design

Abstract

Using a model-following neuro-adaptive approach, a nonlinear controller has been designed in this paper for a practical reusable launch vehicle that assures robust tracking of guidance commands despite having uncertainties in the plant model. The overall control design is carrie d out in two steps. First a nominal design is carried out based on the philosophy of dynamic inversion (for RCS) and optimal dynamic inversion (for aerodynamic control), which assures tracking of the guidance commands for the nominal plant with an optimal control allocation strategy. In this design phase, the bank angle and angle-of-attack commands as issued by the guidance algorithm are first converted to equivalent roll and pitch commands respectively, while simultaneously assuring turn-coordination through the necessary yaw rate command generation. These body rate commands are then tracked in an inner-loop by generating the necessary control surface deflections. Next, an adaptive control is designed that enforces the inner-loop body rates of the actual plant to track the closedloop body rates of the nominal plant. This overall control design structure retains the simplicity in design while simultaneously assuring robust performance of the controller. The control design has been carried out using the full Six-DOF model of the vehicle in velocity frame that imbeds the spherical and rotating earth effects in the vehicle dynamics (which is in harmony with the dynamics used for the vehicle guidance). The promising simulation results with the Six-DOF dynamics along with realistic constraints like actuator dynamics, control bounds, RCS constraints etc. clearly demonstrate the good command following as well as robustness of the overall design approach presented in this paper, making it a viable technique to be implemented in a real vehicle.