Resumo

QUINTERO BERNAL, Daniel Fernando; JARAMILLO BEDOYA, Miguel Ángel; QUINTERO VALLEJO, Hernán Alberto  e  PRADO MARTINEZ, William. Data fusion system for classification of liquefied petroleum gases through machine learning techniques. Ing. Desarro. [online]. 2023, vol.41, n.2, pp.1-.  Epub 08-Nov-2023. ISSN 0122-3461.  https://doi.org/10.14482/inde.41.02.720.891.

Data fusion (DF) is a process that allows the combination of information from various sources for a specific purpose. A DF system, particularly an electronic nose (EN), was designed and built for a device as described in the patent "Portable equipment to avoid vehicle pollution in service stations" with resolution number 23016. It is an instrument for the classification of liquefied petroleum gases (LPG), the application of which is aimed at recognizing diesel fuel or gasolina, in the supply process. These gases are stored in the fuel tank of cars, and by identifying them take the actions required to prevent the contamination of vehicles with a fuel different from that used by the latter. A system that supports the operation of the EN was implemented. In the processing of the information provided by the sensors of the prototype the methodologies of machine learning, K-Nearest Neighbor and Naive Bayes, for lpg differentiation were used. Through a validity test, it was determined that the accuracy of the implemented techniques was 1, therefore, they are ideal methodologies for the classification of diesel fuel and gasoline in dynamic environments.

Palavras-chave : Classification; data fusion; electronic nose; liquefied petroleum gases; machine learning techniques.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )