LPV modeling and gain scheduled bank-to-turn autopilot design

Xun Song,Jianshuang Song,Junbo Wang,Zhang Ren

DOI: https://doi.org/10.1061/41096(366)170

2010-01-01

Abstract:A gain-scheduled bank-to-turn autopilot design for a hypersonic vehicle is presented. The approach using the linear parameter varying transformations doesn't involve linearizations about operating points. Instead, the vehicle dynamics are brought to a linear parameter varying form via a state ransformation. This framework is applied to the design of a coupled longitudinal/lateral bank-to-turn autopilot for a hypersonic vehicle. First, the linear parameter varying (LPV) model of the hypersonic vehicle for bank-to-turn is obtained using state ransformation.The pitch and yaw/roll dynamics are separately transformed to linear parameter varying form, where the cross axis states are treated as exogenous parameters. Because the scheduling variables include these endogenous variables, such a plant is called quasi-linear parameter varying system. Next, using the obtained LPV model, a family of robust controllers is designed for both the pitch and yaw/roll channels, which use angle of attack and roll rate as the scheduling variables. Finally,the closed-loop simulation of the LPV controller is conducted using the original nonlinear model and also using perturbed aerodynamic coefficients. The simulation results show that the desired performance and robustness objectives are achieved across the whole turn phase. © 2010 ASCE.