A constructive approach to stabilizability and stabilization of a class of \(n\)D systems (Q5955550): Difference between revisions

The problem of output feedback stability and stabilization of a class of linear \(n\)D systems (\(n>2\)) is considered. An algebraic approach is proposed by combining Gröbner bases with some existing results on 2D systems. A MIMO \(n\)D system is represented by an \(n\)D rational matrix \(P(z)\) having the matrix fraction description \(P(z)=N(z)D^{-1}(z)\) with \(N(z)\) and \(D(z)\) being \(n\)D polynomial matrices and \(z=(z_{1},\dots,z_{n})\). The reduced minors \(b_{1}(z),\dots,b_{\beta}(z)\) of the system are the maximal order minors of the matrix \([\begin{matrix} N^{T}&D^{T}\end{matrix}]\) divided by their greatest common divisor. The mentioned problem is solved for the class of \(n\)D systems having the following property: the variety of the ideal generated by \(b_{1}(z),\dots,b_{\beta}(z)\) is finite with respect to \(n-2\) variables. The main idea is to decompose this variety into a union of several varieties, each of them being defined by polynomials in just two variables. A stabilizability criterion is given which can be tested constructively, and a procedure to construct a stabilizing compensator is provided. These results are proved for the case \(n=3\) and generalized to the class of \(n\)D (\(n>3\)) systems. An example illustrates the procedure of determining the compensator.