Next-Generation Edge-AI Architectures for Secure EV Charging Infrastructure

Abstract

The growth in the use of electric vehicles (EVs) has led to a massive increase in the demand for infrastructure needed for the charging infrastructure, thus putting an immense amount of pressure on the stability of the grid, cybersecurity and efficiency. In this work, we propose a novel Edge AI architecture for threat and security in the EV charging infrastructure powered by integrating the distributed intelligence, application of real-time anomaly detection and adaptive load optimisation. The proposed system employs artificial intelligence models that are based on fuzzy logic and run at the edge of the network, in the monitoring of the grid conditions, the health of the EVs and cyber threats. Secure coordination between EV's, charging stations and central servers is put in place for increased system reliability, congestion and infrastructure resilience.  

Motivation: Weaknesses in the grid control structures and cybersecurity frameworks have been exposed by the increase in the number of EV charging stations. Uncoordinated charging loads have the potential to destabilise power distribution networks and pose to create complex danger due to cyberattacks and data breaches. Current centralised solutions have latency and scalability issues, which emphasise the need for a decentralised and intelligent framework that can make decisions at the edge level 2 (in real-time) regimes. This paper is motivated by the need to ensure safe, efficient and scalable EV charging operations and protect grid stability and the safety of the critical energy infrastructure.

Novelty: The novelty in this study is to combine Edge AI intelligence with a fuzzy logic-based decision mechanism for the security of EV charging infrastructure. As opposed to typical architectures that are developed on the premise of cloud dependency, the proposed architecture shifts computational intelligence to the charging stations to achieve the specific purpose of ensuring low-latency anomaly detection and adaptive load balancing. The framework provides the unique combination of a grid stress analysis, EV battery health monitoring and cyber security threat analysis as a single decision module. This multidimensional and edge-centric design helps to increase the system resilience and improve the scalability, and allow to correct coordination of the involved components (EV, grid and communication networks).

Findings: Simulation and analytical exams indicate that the proposed Edge-AI architecture has a very positive impact on improving the reliability of charging operations and security of the charging infrastructure. The system can identify the abnormal charging behaviour and counter the cyber threats, and also dynamically adjust the charging loads to prevent grid overloading. Results show a reduction in response time, increased efficiency in load distribution and better security than unauthorised access. The integrated decision framework provides a higher degree of integration of operational stability as opposed to the traditional centralised systems, and this justifies the effectiveness of implementing distributed approaches using Edge AI to change the performance of the EV charging infrastructure and enhance cybersecurity resilience