Controller Design for an Unmanned Reconnaissance Aerial Vehicle
Abstract: The overall objective with this thesis is to evaluate a method for designing nonlinear controllers in aircrafts with concern in robustness against modelling insecurities and also to minimize the design effort. The later objective is of great importance since there exists, in the industry, an ambition to automatize the design for as far extend as possible. The nonlinear method, State Dependent Riccati Equation (SDRE), used in this thesis is a nonlinear version of classic LQ design and both are evaluated and compared for a few flying conditions. Also another nonlinear control method, Two Timescale Separation (TSS), is tested. LQ and SDRE shows equal performance during both looping and more complicated maneuvers, such as high angle of attack wind-vector roll. Further it is possible to automatize the LQ design as well as it is possible for SDRE. Still SDRE is preferable since it will always be somewhat more accurate than LQ. A comparison with the nonlinear method Two Timescale Separation shows results in favor of LQ/SDRE mostly due to relatively slow dynamics and bad accuracy of TSS. A Monte Carlo simulation is made on LQ and SDRE showing that the model is robust against relatively large modelling error of 40%.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)