Single Layer Neural Network Based Precoding and Equalization for 5G Communication System

#massiveMIMO #MIMO #networking #communication #futurenetworks #ieee #ingr

Special Presentation by Dr. Harish Kumar Sahoo (VSSUT, India)

Hosted by the Future Networks Massive MIMO Working Group

Date/Time: Friday, June 6^th, 2025 @ 11:00 am EDT

Topic:

Single Layer Neural Network Based Precoding and Equalization for 5G Communication System

Abstract:

Implementation of hybrid precoding in massive MIMO enhances the spectral efficiency of the receiver in 5G communication system. As hybrid precoding reduces the number of RF chains, the hardware complexity is comparatively less in presence of large number of antenna elements. Moreover, the optimal precoding matrix is helpful to develop efficient equalizer which improves the overall performance of the receiver. Functional link artificial Neural network (FLANN) and recurrent neural network (RNN) can be used as efficient machine learning (ML) models with appropriate training algorithms to generate optimal baseband precoding matrix at the transmitter. FLANN architecture is simpler which consists of a single layer and functional expansion block. Hybrid precoding matrix can be formulated by combining with RF precoding matrix. However, FLANN or RNN based optimization requires reference matrix to carry out the training process which updates the elements of the precoding matrix. Singular value decomposition (SVD) can be one of the approaches to generate the reference matrix through orthogonal decomposition of channel matrix. The research work is focussed on the performance of ML model based precoder for known channel state information (CSI) through simulation by analysing the spectral efficiency with a variation in number of base station (BS) antennas and users within a range of SNR values.

Speaker:

Dr. Harish Kumar Sahoo:  Harish Kumar Sahoo is a senior member of IEEE. He is also a member of IEEE signal processing society, IEEE ComSoc and IEEE Future Networks. He is the current Chair of Technical and Educational Activity Committee IEEE Bhubaneswar Section. He completed his M.Tech from National Institute of Technology, Rourkela, India and Ph.D. from Sambalpur University, India. He is now working as Professor in the Department of Electronics and Telecommunication Engineering, Veer Surendra Sai University of Technology, Burla, Sambalpur, India. He has also additional responsibility of Dean Faculty and Planning of the University.  He has more than twenty-four years of teaching and research experience. Six PhD. Scholars have already been awarded under his guidance. He has several research publications in the journals of IEEE, Elsevier, Springer, Wiley, Taylor and Francis. He is an active reviewer of journal articles for more than thirty-five international publishers including IEEE. His research interests include Massive MIMO, AI/ML Optimization in 5G and 6G communication networks, Channel Estimation and Equalization using adaptive estimation.