AUTOMOTIVE ENGINE DIAGNOSTICS USING FUEL TRIM DATA AND K-NEAREST NEIGHBOR ALGORITHM

Abstract

Technology and science have been revolutionized recently. As a result, the systems on the car are increasingly complicated than it were the past. Thus, it requires new methods to diagnostic the engine and automotive systems’ technical status, rather than depending on the experience of technicans. In this article, I aim to build a model to diagnostic engine status based on Fuel Trims data collected from 300 car samples and using K-nearest Neighbor (KNN) to train this data. The model was successfully built and got the highest accuracy is 87%. The model illustrated the relationship between input data that include age, gender of drivers, the using location of the cars, the milleague of the cars and LTFT index - the index to evaluate the technical status of car engines.