Deep Learning Based Efficient Symbol-Level Precoding Design for MU-MISO Systems

The recently emerged symbol-level precoding (SLP) technique has been regarded as a promising solution in multi-user wireless communication systems, since it can convert harmful interference (MUI) into beneficial signals for enhancing system performance. However, the tremendous computational complexity of conventional designs severely hinders practical implementations. In order to tackle this difficulty, we propose novel deep learning (DL) based approach efficiently design precoders. Particularly, correspondence, consider multi-input single-output (MU-MISO) downlink system. An efficient neural network (EPNN) is introduced optimize precoders maximizing minimum quality-of-service (QoS) all users under power constraint. Simulation results demonstrate that proposed EPNN SLP dramatically reduce computing time at price slight performance loss compared with convex optimization design.