Framework on Deep Learning-Based Joint Hybrid Processing for mmWave Massive MIMO Systems

Dong, Peihao; Zhang, Hua; Li, Geoffrey Ye

doi:10.1109/access.2020.3000601

Cited by 28 publications

(17 citation statements)

References 36 publications

(70 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work outcomes proved the optimal efficiency of energy. A deep learning hybrid processing framework to allow end-to-end optimization by using backpropagation is designed in [18]. The framework includes the design of hybrid processing and signal flow model, which are processed using neural networks.…”

Section: A Prior Workmentioning

confidence: 99%

Performance Enhancement of MmWave MIMO Systems Using Deep Learning Framework

2021

View full text Add to dashboard Cite

In order to obtain beamforming gains and prevent high pathloss in millimeter wave (mmWave) systems, large number of antennas is employed. Digital precoders are difficult to implement with many antennas because of hardware constraints, while analog precoders have limited performance. In this paper, hybrid precoding based on a deep learning framework, HybridPrecodingNet, is proposed, which uses largescale information to predict the parameters of hybrid precoders and decoders. The statistics of the channel covariance matrix are applied to design the hybrid precoders and decoders. The proposed HybridPrecodingNet at the receiver is applied for the channel estimation and design of hybrid decoders. In our proposed framework, the structure of HybridPrecodingNet is trained to learn how to optimize the hybrid precoder and decoder for maximum spectral efficiency. Comparison between different precoding techniques is provided. Results show that HybridPrecodingNet approaches the sub-optimal solution and gives significant spectral efficiency enhancement.

show abstract

Section: A Prior Workmentioning

confidence: 99%

Performance Enhancement of MmWave MIMO Systems Using Deep Learning Framework

2021

View full text Add to dashboard Cite

show abstract

“…Then, the TX designs the hybrid precoder and combiner based on the recovered CSI matrix Ĥ with a DNN. Inspired from the single-timescale deep learning-based hybrid precoding with perfect CSI proposed in [34], we design the hybrid precoding in a two-timescale manner. As shown in Fig.…”

Section: Hybrid Precoder and Combiner Designmentioning

confidence: 99%

“…In this subsection, we introduce how to extend the proposed two-timescale DNN to wideband mmWave OFDM systems. Three key issues need to be considered for the extension [34]:…”

Section: Extension To Ofdm Systemsmentioning

confidence: 99%

See 1 more Smart Citation

Two-Timescale End-to-End Learning for Channel Acquisition and Hybrid Precoding

Hu¹,

Kang²,

Yu³

et al. 2021

Preprint

View full text Add to dashboard Cite

In this paper, we propose an end-to-end deep learning-based joint transceiver design algorithm for millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) systems, which consists of deep neural network (DNN)aided pilot training, channel feedback, and hybrid analog-digital (HAD) precoding. Specifically, we develop a DNN architecture that maps the received pilots into feedback bits at the receiver, and then further maps the feedback bits into the hybrid precoder at the transmitter. To reduce the signaling overhead and channel state information (CSI) mismatch caused by the transmission delay, a two-timescale DNN composed of a long-term DNN and a short-term DNN is developed. The analog precoders are designed by the long-term DNN based on the CSI statistics and updated once in a frame consisting of a number of time slots. In contrast, the digital precoders are optimized by the short-term DNN at each time slot based on the estimated low-dimensional equivalent CSI matrices. A two-timescale training method is also developed for the proposed DNN with a binary layer. We then analyze the generalization ability and signaling overhead for the proposed DNN based algorithm. Simulation results show that our proposed technique significantly outperforms conventional schemes in terms of bit-error rate performance with reduced signaling overhead and shorter pilot sequences.

show abstract

“…Multiple-input-multiple-output (MIMO) in lowfrequency range (e.g., 2-3 GHz) use fully-digital precoding to achieve the sufficient array gain in order to avoid pathloss [6]. To employ the fully-digital precoding, a dedicated radio-frequency (RF) chain is required at each antenna [7]. For large antenna systems (e.g., 256 antennas and above), this traditional fully-digital precoding cannot be used due to the large number of required RF chains as RF chains are powerhungry and expensive [8].…”

mentioning

confidence: 99%

Joint Hybrid Precoder and Combiner for Wideband Millimeter-Wave Massive MIMO Systems

et al. 2020

View full text Add to dashboard Cite

Hybrid precoding and combining is a key technique to provide an appropriate antenna gain in millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems. In wideband mmWave channels, the analog precoder and combiner is designed in the time domain and remain unchanged over the whole bandwidth. In contrast, digital precoders and combiners are optimized on a per-subcarrier basis which makes the resultant problem very difficult. To solve this problem, we combine the well-known turbo-equalizer with the tabu-search (TS)-algorithm developed in artificial intelligence and propose TSbased joint hybrid precoding and combining scheme to intelligently search the near-optimum pair of hybrid precoder and combiner. Specifically, our scheme consists of two key steps. At first, a base station (BS) and a mobile station (MS) develop the turbo-like (TL)-joint search by using the idea of iterative information exchange between them. Then, to find out the near-optimum pair of hybrid precoder and combiner in each iteration of the TL-joint search, the TS-algorithm is employed. Simulation results are shown to verify the significant sum-rate performance of the proposed solution with low-complexity compared to some existing solutions.

show abstract

Framework on Deep Learning-Based Joint Hybrid Processing for mmWave Massive MIMO Systems

Cited by 28 publications

References 36 publications

Performance Enhancement of MmWave MIMO Systems Using Deep Learning Framework

Performance Enhancement of MmWave MIMO Systems Using Deep Learning Framework

Two-Timescale End-to-End Learning for Channel Acquisition and Hybrid Precoding

Joint Hybrid Precoder and Combiner for Wideband Millimeter-Wave Massive MIMO Systems

Contact Info

Product

Resources

About