Rapid prototyping using inversion-based control and object-oriented modelling

Abstract

Object-oriented modelling allows for efficient construction of multi-disciplinary system models in a physically-oriented way. As a unique feature, the modelling approach allows for automatic generation of regular, static, as well as inverse simulation models. Inverse models form the basis of various commonly used nonlinear controller synthesis methods, such as Feedback Linearisation. The resulting multi-variable control laws are easily tuned to meet performance specifications and avoid the need for gain scheduling. In combination with automatic inversion, these synthesis methods are ideally suited for control law rapid-prototyping, allowing experimentation with command variables, control selection and allocation, etc. in very short automated design cycles. After selection of the final architecture, the control laws may be developed further in a detailed design stage. This is demonstrated on the aircraft-on-ground nonlinear control problem for lateral directional control.