Driving Profile Analysis Using Machine Learning Techniques and ECU Data

Automotive vehicles generate a significant amount of data from the engine and its electronic components. The data produced by cars can be used to improve the automotive industry, enhancing driver and vehicle safety, reducing fuel consumption and gas emissions, as well as enabling more complex analyses related to vehicle monitoring by insurers and transport companies, for instance. In this work, we use actual data from an automotive engine ECU as input to machine learning algorithms to analyze the driver’s driving profile and make a correct classification of his/her economy. The algorithms were trained with variables strongly related to fuel consumption, chosen through a feature selection process. Compared to related works, our results are either superior or similar in metrics such as precision, recall, and accuracy, but the algorithms present lower computational costs and real-time analysis through a cloud server.