Condução inteligente: identificação de perfis de condução de veículos pesados para recomendar ações que reduzam o consumo de combustível e emissão de poluentes

Abstract

The transport sector is one of the pillars of a globalized economy with an increasingly efficient logistics chain. Among the modes of transport used to move cargo within Brazil, road transport is the most widely used. With the escalation in the price of fossil fuels and legislation on pollutant gas emissions increasingly strict due to increased concern about global warming, it is crucial that the driving of heavy vehicles powered by combustion engines already produced is optimized. In 2021, to improve its customers’ driving, Volvo do Brasil launched the Volvo Intelligent Driving (CIV) service, which aims to provide fleet managers with a diagnosis of how their vehicles are being driven. This diagnosis consists of a monthly report comparing a certain vehicle’s driving with all those that operated similarly to it. By minimizing fuel consumption and using machine learning algorithms, the service can indicate how well it is driving when compared to the reference (i.e., vehicles in the same application group). This work aims to improve the CIV service using the CAN bus data to give the driver a report of each trip made. This report is available through an application accessible by mobile device, which can be consulted at the end of the journey. With this, the driver will have more control over his driving on specific routes and see the points of improvement to minimize fuel consumption and pollutant emissions. The final solution achieved in this paper shows that it is possible to provide valuable tips about fuel consumption, presented in a user-friendly way, using machine learning algorithms models.