Research Symposium

BIO

Sofia Tenghoff is a freshman majoring in computer science. Her research mentor is Md. Rakibul Ahasan, and she has participated in the national award-winning literary magazine, By Any Other Name. She has attended the machine learning summer program at New York University, and she hopes to become a machine learning engineer and an author after graduating.

Secure EVCS: Cyber-Physical Attack Detection for EV Charging Infrastructure

Abstract

As electric vehicles (EVs) become more common, the need for secure and efficient electric vehicle charging stations (EVCSs) increases. When integrated with smart grid (SG) technology, EVCSs become vital components of smart city infrastructure. However, EVCSs are often subject to cyberattacks, such as manipulations of charging parameters like start time, energy demand, and charging duration by malicious actors. This research project investigates which machine learning (ML) model performs best at detecting cyberattacks in EVCSs. Performance is measured in terms of detection rate (DR). The tested ML models include decision tree (DT), random forest (RF) support vector machines (SVM), and feed forward neural networks (FNN) The purpose of this research is to save EVCS owners time and money and to protect electric vehicle drivers’ data privacy. The methodology involved analyzing data from the ACN website, writing code that builds ML models, and training and testing those models on the data. The results show FNN as the model yielding the best DR.

Keywords: machine learning, cybersecurity, electrical vehicle