Deep Neural Network Assisted Approach for Antenna Selection in Untrusted Relay Networks

This letter mainly studies the transmit antenna selection (TAS) scheme based on deep neural network (DNN) in untrusted relay networks. In our previous work, we revealed that machine learning (ML)-based TAS schemes have performance degradation caused by complicated coupling relationship between the achievable secrecy rate and channel gains. To solve this issue, we here introduce DNN to decouple the above complicated relationship. The simulation results show that the DNN scheme can achieve better decoupling and, thus, accomplish almost the same performance as the exhaustive searching scheme.