Stabilization of TVAR models : A Regularization ApproachM

A new method for the stabilization of time-varying autoregressive models is proposed. The method is based on the interpretation of the underdetermined time-varying least squares prediction problem as an ill-posed inverse problem. The problem is regularized with the Tikhonov approach and is then augmented with nonlinear hyperstability constraints. The problem is solved iteratively with an exterior point algorithm. The performance of the algorithm is studied with a simulation and it is shown that the proposed approach is well suited to stable modeling of burst-like narrow band signals.