Radial-basis-function-based feedforward—feedback control for air—fuel ratio of spark ignition engines

Abstract

In this paper, a feedforward—feedback control is developed for the air—fuel ratio (AFR) of spark ignition engines using neural network estimators. To maintain the AFR at stoichiometric value, the throttle angle change is seen as a disturbance, from which the air flowrate is predicted. The injected fuel is also predicted using the inverse of the fuel injection dynamics. The proposed method is evaluated on an engine simulation benchmark and the performance is shown much improved over proportional—integral control. The new method needs moderate computation and therefore has strong potential to be used in production engines.