Integrated structure/control design based on model validity and robustness margin

Abstract

A design methodology for designing the structure of flexible systems in the presence of modeling uncertainties has been developed. Model validation becomes a critical issue in integrated structure/control design, since the mechanical structure is continuously altered during the iterative design process, and it becomes difficult to guarantee validity of the model used. The proposed method solves this problem with an iterative method by bounding the truncation error as the structural design parameters are changed to optimize the robustness margin of the closed loop system. The method is then applied to the lumped parameter model of a heavy-duty robot with a non-collocated sensor-actuator configuration. This formulation minimizes an objective function over the allowable design parameter subspace based on the structured-singular value of the system. The standard /spl mu/-synthesis methods for controller design tend to provide a very large order controller. It is shown that simultaneous physical structure design and control design can provide an alternative for achieving robust performance without letting the controller order become very high.