2017 年 16 巻 p. 45-53
This paper illustrates a procedure for designing structured robust flight control systems based on Multiple-Delay-Model / Multiple-Design-Point (MDM/MDP) approach, which is one of the useful ways for designing robust flight control systems. The MDM/MDP approach is based on solving a simultaneous stabilization and optimization control problem which minimizes the sum of the stochastic linear quadratic cost indexes weightless on the control inputs. The main contribution of this work is proposing an efficient algorithm for solving the MDM/MDP design problem with less conservatism. Non-convex necessary and sufficient condition of the design problem with static output feedback is derived using an enhanced linear matrix inequality (LMI) characterization. The solution algorithm with guaranteed convergence at least to a local optimal solution is proposed as well, which iteratively solves the LMI sufficient condition of the original non-convex condition by fixing some of the decision variables. This design method is applied to structured roll and side-slip angle controller design. The simulation results show that the proposed method is effective and efficient for robust flight control design with the MDM/MDP approach.