Intelligent hybrid framework for cyber security in EV communication infrastructure

Abstract

Electric Vehicles (EVs) are becoming increasingly popular as a newlineresult of their emphasis on environmentally beneficial modes of transportation, newlinedistributed charging stations, and user-defined supporting infrastructures. newlineElectric vehicle supply equipment (EVSE), which comprises computers with newlineInternet access, is installed in the charging stations so that EVs may be refueled. newlineThese systems are thought to be more crucial for managing tasks like newlineauthorizing and smartly connecting to the local power grid using various newlinewireless technologies like green Wi-Fi, Bluetooth, and even 5G. They also newlineregulate operations like charging electric car batteries. DoS and DDoS attacks newlineare examples of cyberattacks that can compromise the availability, newlineconfidentiality, and integrity of EVSE resources. The focus of recent research newlinehas been on machine and DL algorithm defenses. newlineThe above-mentioned model s implementation in the integrated chip newlinehas grown more challenging since these learning models need big datasets for newlinetraining. The new suggestion to use both learning models with countermeasure newlinestrategies has only been made by a small number of researchers. The system s newlineperformance suffers in terms of factors like prediction ratio and key sensitivity, newlinehowever, because putting the entire architecture on the chip necessitates newlinesignificant computational complexity. newlineThe initial phase of the research suggests a unique intrusion detection newlinearchitecture for EVSE that uses a cat-optimized feed-forward layer to solve the newlineaforementioned issues. In regard to high prediction accuracy (87.13%), newlineprecision (86.88%), recall (87.03%), specificity (87%), and F1-Score newline(86.85%), using the CAT optimization method instead of the conventional newlineELM delivers better results. newline