Deep learning based beamforming for MISO systems with dirty-paper coding

Beamforming technique can effectively improve the spectrum utilization in the multi-antenna systems, while the dirty-paper coding (DPC) technique can reduce the inter-user interference. In this letter, it is aimed to maximize the weighted sum-rate under the total power constraint in the multiple-input-single-output (MISO) system with the DPC technique. However, the existing methods of beamforming optimization mainly rely on customized iterative algorithms, which have high computational complexity. To address this issue, the beamforming neural network (BFNNet) is devised by utilizing the deep learning technique and the uplink-downlink duality and exploring the optimal solution structure, which includes the deep neural network module and the signal processing module. Simulation results show that the BFNNet can achieve near-optimal solutions and significantly reduce computational complexity.

Automated summary

In this research, a beamforming neural network (BFNNet) is developed to maximize the weighted sum-rate in the multiple-input-single-output (MISO) system with dirty-paper coding (DPC) technique. The BFNNet utilizes deep learning, uplink-downlink duality, and explores the optimal solution structure to achieve near-optimal solutions and reduce computational complexity.