Multiobjective Insensitive Computer-Aided Design of Aerospace Control Systems

Abstract A multiobjective CAD method for aerospace control systems has been developed which can meet requirements on disparate objectives over a set of flight conditions, using constrained minimization algorithms with objective functions in the constraint vector. This paper summarizes results of research on four increasingly sophisticated versions of the method: the basic method, an extension which finds Pareto optimal designs which are well-balanced in all objectives, an extension which finds stochastic-insensitive (SI) designs to minimize the sensitivity to uncertain parameters, and a tradeoff method which designs for a compromise between decreased sensitivity and improved nominal objective values. Examples presented are lateral stability augmentation system (SAS) designs to meet multiple handling qualities requirements for a fighter airplane over a range of flight conditions, and similar SAS designs for several flight conditions of the Shuttle entry vehicle. The method requires strong interaction with experienced designers, who must define the objectives and requirements, choose the design flight conditions and the form of the control systems, adjust parameters in the program to get well-balanced designs, and evaluate the design solutions.