Direct Hybrid Precoding in mmWave Massive MIMO Systems

NN_hybrid

The proposed auto-precoder neural network consists of two sections: the channel encoder and the precoder.

Key ideas

  • Learning the channel structure and site-specific channel compression codebooks
  • Leveraging neural networks and prior observations fo fast channel prediction and beam prediction in hybrid analog/digital architectures

Applications

  • Predicting hybrid precoding matrices in mmWave/THz systems with a few pilot signals

More information about this research direction

Paper: X. Li and A. Alkhateeb, “Deep Learning for Direct Hybrid Precoding in Millimeter Wave Massive MIMO Systems,” 2019 53rd Asilomar Conference on Signals, Systems, and Computers, 2019, pp. 800-805.

Abstract: This paper proposes a novel neural network architecture, that we call an auto-precoder. This auto-precoder network jointly senses the millimeter wave (mmWave) channel and designs the hybrid precoding matrices with only a few training pilots. More specifically, the proposed machine learning model leverages the prior observations of the channel to achieve two objectives. First, it optimizes the compressive channel sensing vectors based on the surrounding environment in an unsupervised manner to focus the sensing power on the most promising spatial directions. This is enabled by a novel neural network architecture that accounts for the constraints on the RF chains and models the transmitter/receiver measurement matrices as two complex-valued convolutional layers. Second, the proposed model learns how to construct the RF beamforming vectors of the hybrid architectures directly from the projected channel vector (the received signal). Simulation results show that the proposed approach can significantly reduce the training overhead compared to classical (non-machine learning) solutions. For example, for a system of 64 transmit and 64 receive antennas, with 3 RF chains at both sides, the proposed solution needs only 8 or 16 channel training pilots to directly predict the RF beamforming/combining vectors of the hybrid architectures and achieve near-optimal achievable rates.

@INPROCEEDINGS{Li2019,
author={Li, Xiaofeng and Alkhateeb, Ahmed},
booktitle={2019 53rd Asilomar Conference on Signals, Systems, and Computers},
title={Deep Learning for Direct Hybrid Precoding in Millimeter Wave Massive MIMO Systems},
year={2019},
volume={},
number={},
pages={800-805},
doi={10.1109/IEEECONF44664.2019.9048966}}

To reproduce the results in this paper:

Simulation codes (based on DeepMIMO v2):
These simulations use the DeepMIMO scenarios:
Example: Steps to generate the results in this figure
  1. Download all the files of this GitHub project.
  2. Quick run: Run in terminal “python main_train_beamforming.py -train 1” to train the model and run “python main_train_beamforming.py -train 0” for testing. The default parameters are: dataset=’DeepMIMO_dataset_train20.mat’ and ‘DeepMIMO_dataset_test20.mat’ (which are corresponding to total transmit power of 20dB), epochs=15, batch_size=512, learning_rate=0.002.

Having questions or feedback?

Send an email to 

 

Or post your question in the DeepMIMO forum