ABSTRACT

As will be seen in Part II of this book, tensor product (TP) models with convexity offer great advantages in the process of control system design. It is therefore highly desirable to further transform the higher-order singular value decomposition (HOSVD)-based canonical form of quasi-linear parameter-varying (qLPV) models to become TP model incorporating various convexities. The convexity induced will in turn define the type of convex hull produced by the linear time-invariant (LTI) vertex systems of the resulting TP model, and yield strong influence on the control design feasibility and performance. This chapter serves to present a systematic way of generating different types of convex TP models.