A Novel Deep Neural Network Based Antenna Selection Architecture for Spatial Modulation Systems

With the constantly developing technology, the speed and accuracy requirement of communication systems are increasing day by day. Spatial modulation (SM) is a recent and promising technique which additionally uses antenna indices for multiple input multiple output (MIMO) systems. In order to add another degree of freedom to SM's efficiency, transmit antenna selection (TAS) algorithms are a crucial field to study. On the other hand, use of artificial intelligence significantly developed in nowadays in wide variety of areas such as biology, robotics, automation etc. The main purpose of this study is to realize TAS for SM systems using deep neural network (DNN). Besides, the processing load of the proposed DNN is reduced without involving the repetitive parts of the TAS metric which is not studied in the literature as far as we know. It is shown that the proposed DNN based TAS algorithm outperforms existing studies in terms of symbol error rate.

Automated summary

This study utilizes deep neural networks to implement transmit antenna selection for spatial modulation systems, showing superior performance in terms of symbol error rate compared to existing studies.