Automatic control techniques for parameters related to the air–fuel ratio in internal combustion engines: A review

Abstract

Global regulations on vehicle emissions have compelled the automotive industry to undergo rapid advancements over the past decades, prioritizing strategies for developing engines with greater efficiency and lower emissions. While the market for electric vehicles continues to expand, internal combustion (IC) engines remain indispensable and constitute the majority of the global vehicle fleet. Among the critical parameters for controlling engine performance, emissions, and fuel economy, the air–fuel ratio (AFR) is a key determinant. Inadequate control or fluctuations in AFR, whether in steady-state or transient conditions, can lead to higher levels of pollutant emissions, increased fuel consumption, and engine instability. For more than a century, AFR control has progressed from carburetor-based systems to electronically controlled port fuel injection (PFI) or gasoline direct injection (GDI) systems. In PFI or GDI systems, the precision of AFR regulation depends heavily on the performance of the controller governing the actuation of fuel injectors. This article reviews recent advancements in AFR control techniques, categorizing them into feedback, model-based, parameter estimation, robust, fuzzy, adaptive, neural network (NN), and machine learning (ML) approaches. At the end of each section, a comparative discussion outlines the most appropriate use cases for each technique, their relative performance against alternative approaches, implementation limitations, and studies that have demonstrated their applicability in real-world scenarios. Furthermore, graphical representations are provided to illustrate the annual volume of published research on IC engine control using each technique. These figures offer valuable insights into prevailing research trends and support the identification of promising directions for the development and implementation of future control strategies.