Optimal Control of EGR System in Gasoline Engine Based on Gaussian Process

Abstract

The contribution described in this paper is concentrated on the integration of exhaust gas recirculation (EGR) system into the process of combustion in an optimal manner. In practice, deriving a state-space model of this actuator is an energetic task as a result of involving some uncertain chemical reactions. To alleviate the effect of unobserved phenomena, which does not seem to be easy in modeling, an improved Gaussian Process (GP) is represented for identifying such dynamics. In this approach, practical modification in general formulation of GP is provided based on proportional feedback gain adjustment. Afterwards, the obtained model is considered for design of optimal model-based control strategy. The whole aim is focused on achieving a green economically gasoline engine by optimizing the trend of fuel consumption. Eventually, simulation results illustrate the effectiveness of proposed structure in EGR systems.