Nonlinear Optimal Generalized Predictive Functional Control Applied to Quasi-LPV Model of Automotive Electronic Throttle

Abstract

A Nonlinear Optimal Generalized Predictive Functional Control algorithm is presented for the control of quasi linear parameter varying state-space systems. A scalar automotive electronic throttle body is simulated to demonstrate typical results. The controller structure is specified in a restricted structure form including a set of pre-specified linear transfer-functions and a vector of gains that are found to minimize a GPC cost-index. This approach enables a range of classical controller structures to be used in the feedback loop such as extended PI, PID or of a more general transfer-function form. The controller is introduced along with a dynamic cost-weighting tuning future. A simulation is used to validate the performance of the restricted structure controller for regulation and tracking problems assessed against automotive performance standards.