Optimal decoupling for MIMO-controller design with robust performance

MIMO-identification and robust controller design are cumbersome. As a result, MIMO-systems are often controlled by decentralized control systems, which consist of independent SISO-controllers based on the diagonal elements of the system. The neglected off-diagonal elements limit however the performance. This paper presents a design approach that combines decentralized control design with an input/output decoupling transformation yielding higher closed loop performance. This approach consists of a procedure to find the transformations of the inputs and the outputs such that the relation between the transformed inputs and outputs is as diagonal as possible. Then, decentralized control techniques are used to design independent SISO-controllers for the optimal decoupled system, which guarantee robust performance of the overall system. The complexity of this combination is comparable to that of a decentralized controller, but the performance approaches that of a full MIMO-controller. The optimal decoupling quality, and accordingly, the achievable closed loop performance depend on the symmetry in the system. Validation results on a automotive durability test rig simulation shows that the controller designed with decoupling yields better performance than a decentralized controller.