Source author record

Mansi Girdhar

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

Cryptography and Security eess.SY Machine Learning Systems and Control

Catalog footprint

Research overview

2works

4topics

4close collaborators

Actions

Is this your work?

Open graph Browse papers

preprint2022arXiv

ECU Identification using Neural Network Classification and Hyperparameter Tuning

Intrusion detection for Controller Area Network (CAN) protocol requires modern methods in order to compete with other electrical architectures. Fingerprint Intrusion Detection Systems (IDS) provide a promising new approach to solve this problem. By characterizing network traffic from known ECUs, hazardous messages can be discriminated. In this article, a modified version of Fingerprint IDS is employed utilizing both step response and spectral characterization of network traffic via neural network training. With the addition of feature set reduction and hyperparameter tuning, this method accomplishes a 99.4% detection rate of trusted ECU traffic.

preprint2022arXiv

Machine Learning-Enabled Cyber Attack Prediction and Mitigation for EV Charging Stations

Safe and reliable electric vehicle charging stations (EVCSs) have become imperative in an intelligent transportation infrastructure. Over the years, there has been a rapid increase in the deployment of EVCSs to address the upsurging charging demands. However, advances in information and communication technologies (ICT) have rendered this cyber-physical system (CPS) vulnerable to suffering cyber threats, thereby destabilizing the charging ecosystem and even the entire electric grid infrastructure. This paper develops an advanced cybersecurity framework, where STRIDE threat modeling is used to identify potential vulnerabilities in an EVCS. Further, the weighted attack defense tree approach is employed to create multiple attack scenarios, followed by developing Hidden Markov Model (HMM) and Partially Observable Monte-Carlo Planning (POMCP) algorithms for modeling the security attacks. Also, potential mitigation strategies are suggested for the identified threats.

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.