Transformer learning-based efficient MIMO detection method

Signal detection for multiple-input-multiple-output (MIMO) systems is a challenging problem due to its computational complexity. The conventional algorithms used in this problem often are either impractical or suffer from performance limitations. In this paper, we propose a machine learning-based MIMO detection method. The proposed method employs the encoder block of a transformer learning approach that has been tailored for MIMO detection. The input to the network of the proposed method is prepossessed using a simple linear decomposition method. Simulation results show that the proposed method achieves a significant enhancement in bit error rate (BER) performance and ultimately produces performance approaching that of the maximum likelihood (ML) detection method.