Content area
Abstract
In H∞ control design two major factors determine the success of the design in achieving performance and robustness requirements. These factors are the cost function which represents the objectives and the weighting functions which are used to achieve the objectives. In this thesis a generalised cost function is proposed which captures important features of the feedback system. Furthermore, attempts are made to present new ideas on the issue of weighting function selection which seems to be a problem dependent issue. A new perception of the weighting functions is presented to help in choosing the weighting functions together with many general rules and guidelines. This thesis also presents two new methods by which the problem caused by having a restricted number of sensors could be overcome. These simple methods are suitable for H∞ control optimisation framework, and they rely on transforming the output variables domains.