Battery Range Estimation In Electric Vehicles Using Machine Learning and Deep Learning Techniques

Abstract

Greenhouse gas emission has been a prominent issue leading to global warming. One of the reasons is the use of motor vehicles fueled by oil and natural gas. These vehicles burn the oil and gases to power their engines that leads to the release of several gases into the Earth's atmosphere. One of these gases is CO2, a greenhouse gas contributing to global warming. Electric vehicles popularly known as EVs becoming a major solution to this issue. However, they face few practical issues like the long wait time to recharge the batteries. One of the issues is the estimation of remaining distance coverage with the available battery power. This is crucial as it helps the users to plan their next recharge efficiently. This paper presents the application of machine learning in estimating the travel distance that can be covered with the available battery power remaining. This paper explored several state-of-the-art machine learning models and propose LSTM as the suitable technique to accurately estimate the distance that can be covered with the remaining battery power. The models were evaluated using an open-source dataset taken from Kaggle. The proposed model achieved a mean R2 score of 0.84 and demonstrated the potential of LSTM technique for this problem. The findings of this work is time relevant as the demand for EVs are increasing everyday. © 2024 IEEE.