Polytopic Decomposition of the Linear Parameter-Varying Model Based on HOSVD

For the solution of an infinite number of LMI in the parameters trajectory of LPV system, a method based on the tensor product transformation is proposed to tranform the LPV system into the convex polytopic structure. Firstly, discretizing the given LPV system within the interzone of the changing parameters and storing it into a tensor, and then using higher order singular value decomposition (HOSVD), discarding smaller and non-zero singular values and their corresponding singular vectors, reconstructing the reduced-rank tensor to obtain a finite number of LTI vertex systems. The final example shows that the convex polytopic system obtained by this method can substitute the original LPV system within the allowed error.