Robust decentralised control design for a MIMO nonminimum-phase process

The design of stabilising diagonal controllers for linear highly-coupled multiple-input multiple-output systems is presented. An alternative approach to the relative gain array method for channel-coupling evaluation is used. It is based on the individual channel analysis and design. This approach takes advantage of the appealing properties of the multivariable structure function (MSF). This framework is shown to be effective in the analysis and control of a well-known quadrupletank process, which is nonminimum phase for some operating points. Indices of performance, cross-coupling between inputoutput channels, structural stability and robustness are established by means of the appropriate interpretation of the MSF. Frequency-domain-based techniques are used. The MSF gives a clear and systematic design process for both input-output channels. Furthermore, diagonal stabilising controllers for the two possible input-output pairings are obtained. Both pairings give rise to open-loop nonminimum-phase systems. The main contribution is the proof that one of these pairings renders the closed-loop system minimum phase through an appropriate control design