With the development of vehicle intelligent technology, the combination of network and vehicle has become an inevitable trend, which brings great convenience to people. Meanwhile, hackers can also use technical loopholes to hijack vehicles, leading to serious traffic accidents and vehicle crashes. These accidents of vehicles network security have caused huge economic losses to enterprises and governments. Traditional single-point protection technology is lack of effective cooperation, and cannot hold the network security status of the Internet of Vehicles as a whole. To face of infinite attacks on the Internet of Vehicles network, the Internet of Vehicles needs situational awareness technology to protect it. This research applies to federated learning to the Internet of vehicles. Traditional machine learning trains the model on a single machine. Distributed learning needs to collect and allocate the data, which increases the transmission delay and bandwidth of the data, and has higher requirements on the transmission speed and bandwidth of the network. Federated learning combines data on different clients to jointly train on a common model without data sharing, which protects the privacy of users. Through experiments, it is proved that the use of federated learning algorithm in the prediction of the security situation of the Internet of Vehicles network has a high accuracy rate for the prediction of attacks.
Research on Internet of Vehicles Attack Prediction Based on Federated Learning
2021-12-10
1030586 byte
Conference paper
Electronic Resource
English
METHODS TO IMPROVE FEDERATED LEARNING ROBUSTNESS IN INTERNET OF VEHICLES
European Patent Office | 2024
|Methods to Improve Federated Learning Robustness in Internet of Vehicles
European Patent Office | 2024
|