High Performance multivariable spectral estimator with Bounded McMillan Degree

Spectral estimation is a prominent issue with applications in a widespread variety of fields, from signal processing to biology. To the present, multivariable spectral estimation with complexity constraints on the acceptable solutions is a challenging issue. In this work, this subject is recast in the form of a constrained optimization problem. This can be efficiently tackled by means of duality theory, because the corresponding dual problem turns out to be particularly suitable for a solving strategy based on Newton-type algorithmic approach. Existence and uniqueness of the solution of the dual problem are proven, and the global convergence of the proposed algorithm is proved. Numerical simulations suggest the efficiency of the proposed technique