A NEURAL NETWORK AIR-FUEL RATIO ESTIMATOR FOR CONTROL AND DIAGNOSTICS IN SPARK-IGNITED ENGINES

The paper focuses on the experimental identification and validation of a recurrent neural network (RNN) estimator of air-fuel ratio (AFR) excursions in sparkignited engines. Suited training procedures and experimental tests are proposed to improve RNN precision and generalization in predicting AFR transients for a wide range of operating scenarios. The reference engine has been tested by means of an integrated system of hardware and software tools for engine test automation and control strategies prototyping. The simulation performed on the test-sets show the ability of the RNN to reproduce the target patterns with satisfactory accuracy. The results achieved envisage potential applications of the AFR estimator in real-time diagnostics and control algorithms.