Yin Yujia, Fang He
School of Electronic and Information Engineering, Soochow University, Suzhou 215301, China.
Sensors (Basel). 2023 Jul 28;23(15):6751. doi: 10.3390/s23156751.
Six-generation (6G) networks will contain a higher density of users, base stations, and communication equipment, which poses a significant challenge to secure communications and collaborations due to the complex network and environment as well as the number of resource-constraint devices used. Trust evaluation is the basis for secure communications and collaborations, providing an access criterion for interconnecting different nodes. Without a trust evaluation mechanism, the risk of cyberattacks on 6G networks will be greatly increased, which will eventually lead to the failure of network collaboration. For the sake of performing a comprehensive evaluation of nodes, this paper proposes a novel multiple role fusion trust evaluation framework that integrates multiple role fusion trust calculation and blockchain-based trust management. In order to take advantage of fused trust values for trust prediction, a neural network fitting method is utilized in the paper. This work further optimizes the traditional trust management framework and utilizes the optimized model for node trust prediction to better increase the security of communication systems. The results show that multiple role fusion has better stability than a single role evaluation network and better performance in anomaly detection and evaluation accuracy.
第六代（6G）网络将包含更高密度的用户、基站和通信设备，由于网络和环境复杂以及所使用的资源受限设备数量众多，这对安全通信与协作构成了重大挑战。信任评估是安全通信与协作的基础，为不同节点的互连提供了接入标准。没有信任评估机制，6G网络遭受网络攻击的风险将大大增加，最终会导致网络协作失败。为了对节点进行全面评估，本文提出了一种新颖的多角色融合信任评估框架，该框架集成了多角色融合信任计算和基于区块链的信任管理。为了利用融合后的信任值进行信任预测，本文采用了一种神经网络拟合方法。这项工作进一步优化了传统的信任管理框架，并利用优化后的模型进行节点信任预测，以更好地提高通信系统的安全性。结果表明，多角色融合比单角色评估网络具有更好的稳定性，在异常检测和评估准确性方面具有更好的性能。
