Driving Range Estimation and Energy Consumption Rate Deviation Classification in Electric Vehicles using Machine Learning Methods
Using a dataset collected from
https://spritmonitor.de/
driving range for electric vehicles
is predicted via input features, such as driving_style
, avg_speed
and route_type
.
- Regressors:
- Linear Regression
- Multilayer Perceptron (MLP)
- Random Forest
- AdaBoost
- Deep Multilayer Perceptron (Deep MLP)
- Classifiers:
- Support Vector Machines (SVM)
- Multilayer Perceptron (MLP)
- Random Forest
- Deep Multilayer Perceptron (Deep MLP)
Find the related published conference paper here.
@inproceedings{amirkhani2019electric,
title={Electric Vehicles Driving Range and Energy Consumption Investigation: A Comparative
Study of Machine Learning Techniques},
author={Amirkhani, Abdollah and Haghanifar, Arman and Mosavi, Mohammad R},
booktitle={2019 5th Iranian Conference on Signal Processing and Intelligent Systems (ICSPIS)},
pages={1--6},
year={2019},
organization={IEEE}
}
Dataset crawler (vehicle_crawler.py
) and
an example result (volkswagen_e_golf.csv
) in csv file can be found here:
https://github.com/armiro/crawlers/tree/master/SpritMonitor-Crawler
First, change the dataset path in both files. Then,
- run the
driving_range_prediction.py
file to predict the trip distance of the electric vehicle; how long this vehicle can go in the next trip. - run the
ECR_deviation_classification.py
file to classify the ECR deviation from the manufacturer; whether in this trip ECR is higher (more consumption) or lower (less consumption) than the factory-defined ECR.