Transmit antenna selection for spatial modulation based on machine learning

The next-generation of networks should have better spectrum and energy efficiency. These competing design parameters can be dealt with spatial modulation (SM). Transmit antenna selection (TAS) strategies can enhance the average bit error rate (ABER) performance of SM. The purpose of this research is to examine the employment of machine learning (ML)-based TAS schemes for SM. The conventional Euclidean distance-based antenna selection (EDAS) scheme offers optimal performance at the expense of computational complexity. This paper discusses the application of four different ML algorithms to SM-TAS: Support vector machines (SVM), K-nearest neighbour (KNN), decision tree (DT), and naïve Bayes (NB). With reduced complexity, an SVM-based TAS scheme for SM can offer a minimum gain of ∼3.4 dB over SM with no transmit antenna selection (SM-NTAS). Furthermore, the ABER performance of all schemes is tested with a higher modulation order and a number of transmit antennas. In every scenario, ML-based TAS schemes outperform SM-NTAS in terms of ABER performance.