A Self-Learning Detection Method of Sybil Attack Based on LSTM for Electric Vehicles

Abstract

Electric vehicles (EVs) are the development direction of new energy vehicles in the future. As an important part of the Internet of things (IOT) communication network, the charging pile is also facing severe challenges in information security. At present, most detection methods need a lot of prophetic data and too much human intervention, so they cannot do anything about unknown attacks. In this paper, a self-learning-based attack detection method is proposed, which makes training and prediction a closed-loop system according to a large number of false information packets broadcast to the communication network. Using long short-term memory (LSTM) neural network training to obtain the characteristics of traffic data changes in the time dimension, the unknown malicious behavior characteristics are self-extracted and self-learning, improving the detection efficiency and quality. In this paper, we take the Sybil attack in the car network as an example. The simulation results show that the proposed method can detect the Sybil early attack quickly and accurately.