| The main goal of this work is to design a control law for a commercial airplane autopilot, available through the entire flight envelope. The first step of such a project is to obtain the linearised model of the system we seek to control. This was carried out based on a full compendium edited by the Boeing Company Inc. of aerodynamic data and simulations of the B-747 aircraft we had at our disposal. Firstly, we obtained its nonlinear model and then, after decoupling, its linearised longitudinal model which is available for any flight condition.; Secondly, after a review of command law synthesis methods for aircraft, we decided to design a control structure containing several controllers which switch from one to another according to the flight condition of the aircraft. The method used, whose theory is widely studied, was the H∞ synthesis which gives the necessary robustness to controllers to perform well over a wide enough flight domain. After studying and validating a method to ensure bumpless switching between command signals (the so-called observer-based method), we studied a control structure with gain-scheduling. A performance criterion was chosen to select which controllers had to be used in this structure, considering that the aim was to use the least number of them.; The last step consisted in validating the designed control law: we sought to ensure its smooth functioning and to reach the performance awaited. Moreover, this work led to designing graphical interfaces permitting to easily apply the results presented in this dissertation, and to show the motions of the Boeing 747 through an animation. |
