| Title |
Multi-task Learning-deep Neural Network-based Secrecy Rate Maximization for Multiple Intelligent Reflecting Surface System |
| Authors |
문상미(Sangmi Moon) ; 황인태(Intae Hwang) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.10.19 |
| Keywords |
Deep neural network; Intelligent reflecting surface; Multi-task learning; Secrecy rate |
| Abstract |
In this paper, we propose deep learning scheme-based secure transmission in multiple intelligent reflecting surface (IRS) millimeter-wave system. The proposed scheme predicts active IRS and phase shift based on multi-task learning in deep neural network to maximize the secrecy rate. Simulation results based on 3D ray-tracing show that proposed scheme could predict the active IRS and phase shift with an accuracy exceeding 96%. In addition, the proposed scheme has a higher secrecy rate than the conventional single IRS and multiple IRS. |