Energy Prediction for EVs Using Support Vector Regression Methods

This paper presents the application of machine learning algorithms for an accurate estimation of the energy consumption of electric vehicles (EVs). Normalised energy consumption values and speed profiles are collected from various EVs for a cloud-based prediction approach. We predict the necessary energy for each road segment on the basis of crowd-sourced data. Support vector machines, which are trained by the collected historical data of the driver, predict the deviation from the average energy consumption on each road segment. As a result, the prediction of propulsion energy consumption for EVs before the start of a trip has a relative mean error of less than 6.7%.     «

This paper presents the application of machine learning algorithms for an accurate estimation of the energy consumption of electric vehicles (EVs). Normalised energy consumption values and speed profiles are collected from various EVs for a cloud-based prediction approach. We predict the necessary energy for each road segment on the basis of crowd-sourced data. Support vector machines, which are trained by the collected historical data of the driver, predict the deviation from the average energy...     »