NGMV control of state dependent multivariable systems

Abstract

A nonlinear generalized minimum variance control law is derived for the control of nonlinear multivariable systems. The solution for the control law is obtained using a model for the process that includes three different types of subsystem to provide a variety of means of modeling the system. These may not all be present. The first subsystem involves a nonlinear operator of very general form. The second is a state dependent, state equation model of the plant and finally a linear state-equation model represents the output subsystem, disturbance and reference models. The process is assumed to include common delays in input or output channels of magnitude k. The quadratic like cost index involves state, error and control signal costing terms. The controller obtained is simple to implement, particularly in one form, which might be considered a nonlinear state-dependent version of the Smith predictor.