Electric Vehicles Driving Range and Energy Consumption Investigation: A Comparative Study of Machine Learning Techniques

Electric vehicles are the next generation of cars which are pollutant-free, resulting in the elimination of many environmental and healthcare problems caused by fossil-fueled vehicles. On the other hand, mass production and wide adoption of these vehicles are facing significant barriers; long battery charging time and limited trip distance per charge are the most important ones to mention. Due to the development of fast DC chargers, the former problem is resolved to a certain extent, while the latter is still a topic of interest. In this article, using a publicly available dataset, driving range estimation of a specific electric vehicle model is scrutinized. At first, multiple regression models are trained based on the features like the average speed, type of the route and driving style; then the driving range prediction accuracy is investigated. On the next step, sensitivity analysis is performed on the energy consumption rate, and the results are discussed. At the final phase, the effect of each feature on the energy consumption rate is highlighted, and the deviation between experimental rates and the factory-defined rates are explained in details.