Signal detection based on deep learning in MIMO-OFDM systems

The signal detection is presented based on deep learning (DL) for Multiple-in-Multiple-out Orthogonal Frequency Division Multiplexing (MIMO-OFDM) Systems. In MIMO-OFDM systems, a receiver is designed to eliminate successive interference cancellation based on DL for multiple users. The signal detection and channel estimation are carried out using deep neural network (DNN) which is trained offline depend on simulation data. And the symbols online are recovered directly. The simulation results show that Deep learning (DL) method is better than those traditional methods for channel estimation. The error propagation effects are reduced by DNN in the signal detector. The inter-symbol interference (ISI) of systems is serious, which shows that the DL approach can achieve the better performance by the DL approach than the maximum likelihood approach.