Electric Vehicle Range Prediction

Abstract

The introduction of new energy vehicles has emerged as a new trend in the automotive industry in
response to growing energy and environmental issues. The electric vehicle (EV) is the driving force behind new
energy vehicles. The one major problem electric vehicles have always been the distance(range) of the travel and map
to the nearby charging stations. For range prediction in the present study, four machine learning
algorithms—multiple linear regression, random forest regression, polynomial regression, and support vector
regression—are employed and contrasted with the most recent research. Python implementation is used for a
simulation model that takes the coordinates of the starting point, endpoint, Map of the surrounding location land, and
the locations of charging stations to predict an optimal path for the vehicle to travel. We use an A* path-finding
algorithm to traverse the black-and-white map and find the optimal path from source to destination. This work
proposes a novel Machine learning (ML) based predictive strategy to estimate the driving range of electric vehicles
and route them to the nearby charging stations. The range predictor considers the specific vehicle parameters over a
distributed network of charging stations. The charging stations are modeled as entities with many charging points
and, in each stage of the prediction the status of availability of charging points is monitored and updated to a cloud
database. A better overview of the driving range and vehicle’s energy consumption may help reduce the overall range
anxiety of many EV drivers.