Generalized predictive control applied to DC-DC converters

This paper focuses on the application of a generalized predictive control (GPC) law design technique for digitally controlled DC-DC converters. The unconstrained GPC control law presented is based on a minimization of a 2-norm measure of predicted performance. It provides an automatic method of designing digital linear compensators based on discrete representations of the plant model. The key characteristic of this approach is that for specific design criteria, the GPC designed compensator always guarantees stability. This provides a very useful starting point to develop and optimize the compensator for desired closed-loop performance. For illustration purposes, simulation results are presented for a standard second-order transfer function model of a Buck converter.