Sub-Array Hybrid Precoding for Massive MIMO Systems: A CNN-Based Approach
Published in Journal, 2021
Recommended citation: K. Chen, J. Yang, Q. Li and X. Ge, "Sub-Array Hybrid Precoding for Massive MIMO Systems: A CNN-Based Approach," in IEEE Communications Letters, vol. 25, no. 1, pp. 191-195, Jan. 2021, doi: 10.1109/LCOMM.2020.3022898. https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9189869
In order to reduce the computation time of hybrid precoding processing while improving the spectrum efficiency (SE) of massive multiple-input multiple-output (MIMO) systems, in this letter, we investigate the sub-array hybrid precoding based on the convolutional neural network (CNN). A constraint-relaxation alternating minimization (CR Alt-Min) algorithm is proposed to create the training set of the CNN. To reduce the computation time caused by iterations in the Alt-Min algorithm, a CNN-based algorithm is proposed. Simulation results show that the CNN-based algorithm reduces the computation time in hybrid precoding processing by an order of magnitude. Moreover, the maximum SE is improved by 26.64% by the CNN-based algorithm, compared with the Alt-Min algorithm.
Recommended citation: K. Chen, J. Yang, Q. Li and X. Ge, “Sub-Array Hybrid Precoding for Massive MIMO Systems: A CNN-Based Approach,” in IEEE Communications Letters, vol. 25, no. 1, pp. 191-195, Jan. 2021, doi: 10.1109/LCOMM.2020.3022898.