Multiuser physical layer security mechanism in the wireless communication system of the IIOT

Wireless system in industrial scene plays an important role in the process of automation. This kind of system urgently needs low complexity, lightweight, high security authentication mechanism. The emergence of physical layer authentication meets these requirements. However, the existing authentication mechanism based on binary hypothesis testing can only perform ideally under fixed conditions, and cannot distinguish multiple users; The authentication mechanism based on deep neural network (DNN) algorithm has limitations in small sample learning and parameter setting. In order to further improve the accuracy of authentication in dynamic industrial scenarios, a new multiuser physical layer authentication scheme is proposed. The mechanism uses machine learning algorithm based on autonomous parameter optimization to replace the traditional decision making method based on user-defined threshold, and is suitable for small sample learning. This paper takes the channel matrix estimated by the mobile node as the authentication input, obtains different channel matrix dimensions through down sampling, and finds out the optimal channel matrix dimension through experiments, so as to reduce the running time and improve the authentication accuracy. A large number of simulations are carried out using the public dynamic industrial scene data set. Compared with the existing authentication schemes, the proposed authentication scheme further improves the accuracy of multiuser authentication in dynamic industrial scenarios.