Deep Learning Based Massive-MIMO Decoder

Kumar, Satish; Singh, Anurag; Mahapatra, Rabi

doi:10.1109/ants47819.2019.9118152

Cited by 5 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DLNet [17], [30] • A new loss function is proposed based on the sum of mean squared error between the transmitted and estimated signals at each layer.…”

Section: Technique Advantages Disadvantagesmentioning

confidence: 99%

Massive MIMO Detectors Based on Deep Learning, Stair Matrix, and Approximate Matrix Inversion Methods

Albreem,

Alnajjar,

Almasadeh

2023

IEEE Access

View full text Add to dashboard Cite

Massive multiple-input multiple-output (MIMO) is an essential technology in fifth-generation (5G) and beyond 5G (B5G) communication systems. Massive MIMO is employed to meet the increasing request for high capacity in next-generation wireless communication networks. However, signal processing in massive MIMO incurs a high complexity due to a large number of transmitting and receiving antenna elements. In this paper, we propose low complexity massive MIMO data detection techniques based on zero-forcing (ZF) and vertical bell laboratories layered space-time (V-BLAST) method in combination with approximate matrix inversion techniques; Neumann series (NS) and Newton iteration (NI). The proposed techniques reduce the complexity of the ZF V-BLAST method since they avoid the exact matrix inverse computation. Initialization based on a stair matrix is also exploited to balance the performance and the complexity. In addition, we propose a massive MIMO detector based on approximate matrix inversion with a stair matrix initialization and deep learning (DL) based detector; MM Network (MMNet) algorithm. MMNet contains a linear transformation followed by a non-linear denoising stage. As signals propagate through the MMNet, the noise distribution at the input of the denoiser stages approaches a Gaussian distribution, form precisely the conditions in which the denoisers can attenuate noise maximally. We validated the performance of the proposed massive MIMO detection schemes in Gaussian and realistic channel models, i.e., Quadriga channels models. Simulations demonstrate that the proposed detectors achieve a remarkable improvement in the performance with a notable computational complexity reduction when compared to conventional ZF V-BLAST and the MMNET in both simple and real channel scenarios.

show abstract

“…DLNet [17], [30] • A new loss function is proposed based on the sum of mean squared error between the transmitted and estimated signals at each layer.…”

Section: Technique Advantages Disadvantagesmentioning

confidence: 99%

Massive MIMO Detectors Based on Deep Learning, Stair Matrix, and Approximate Matrix Inversion Methods

Albreem,

Alnajjar,

Almasadeh

2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…In [165], a DL based network (DLNet) for signal decoding in massive MIMO systems is proposed. It considers timevarying Gaussian random channel with a perfect CSI at the receiver.…”

Section: Belief Propagation (Bp)mentioning

confidence: 99%

“…• It offers a better BER performance with 164× faster in running speed and 9× less computational complexity than the DetNet architecture [165]. • It achieves a comparable BER performance with the SDR algorithm, with 28200× faster [165].…”

Section: Dlnetmentioning

confidence: 99%

Deep Learning for Massive MIMO Uplink Detectors

Albreem

Alhabbash²,

Shahabuddin

et al. 2022

IEEE Commun. Surv. Tutorials

View full text Add to dashboard Cite

Detection techniques for massive multiple-input multiple-output (MIMO) have gained a lot of attention in both academia and industry. Detection techniques have a significant impact on the massive MIMO receivers' performance and complexity. Although a plethora of research is conducted using the classical detection theory and techniques, the performance is deteriorated when the ratio between the numbers of antennas and users is relatively small. In addition, most of classical detection techniques are suffering from severe performance loss and/or high computational complexity in real channel scenarios. Therefore, there is a significant room for fundamental research contributions in data detection based on the deep learning (DL) approach. DL architectures can be exploited to provide optimal performance with similar complexity of conventional detection techniques. This paper aims to provide insights on DL based detectors to a generalist of wireless communications. We garner the DL based massive MIMO detectors and classify them so that a reader can find the differences between various architectures with a wider range of potential solutions and variations. In this paper, we discuss the performance-complexity profile, pros and cons, and implementation stiffness of each DL based detector's architecture. Detection in cell-free massive MIMO is also presented. Challenges and our perspectives for future research directions are also discussed. This article is not meant to be a survey of a mature-subject, but rather serve as a catalyst to encourage more DL research in massive MIMO.

show abstract

“…In addition, DL can also be integrated into the encoder/decoder for detection problems. For instance, [204] proposes a DLNet decoder to minimize the bit error rate. In [205], a concrete map detection (CMD) is proposed, which relaxes the probability mass function of the discrete random variable into a probability density function in a maximum a posterior detection problem.…”

Section: F I-mmimo Decodermentioning

confidence: 99%

Intelligent Massive MIMO Systems for Beyond 5G Networks: An Overview and Future Trends

et al. 2022

View full text Add to dashboard Cite

Machine learning (ML) which is a subset of artificial intelligence is expected to unlock the potential of challenging large-scale problems in conventional massive multiple-input-multiple-output MIMO (CM-MIMO) systems. This introduces the concept of intelligent mMIMO (I-mMIMO) systems. Due to the surge of application of different ML techniques in the enhancement of mMIMO systems for existing and emerging use cases beyond fifth-generation (B5G) networks, this article aims to provide an overview of the different aspects of the I-mMIMO systems. First, the characteristics and challenges of the CM-MIMO have been identified. Secondly, the most recent efforts aimed at applying ML to a different aspect of CM-MIMO systems are presented. Thirdly, the deployment of I-mMIMO and efforts towards standardization are discussed. Lastly, the future trends of I-mMIMO enabled application systems are presented. The aim of this paper is to assist the readers to understand different ML approaches in CM-MIMO systems, explore some of the advantages and disadvantages, identify some of the open issues, and motivate the readers toward future trends.

show abstract

Deep Learning Based Massive-MIMO Decoder

Cited by 5 publications

References 12 publications

Massive MIMO Detectors Based on Deep Learning, Stair Matrix, and Approximate Matrix Inversion Methods

Massive MIMO Detectors Based on Deep Learning, Stair Matrix, and Approximate Matrix Inversion Methods

Deep Learning for Massive MIMO Uplink Detectors

Intelligent Massive MIMO Systems for Beyond 5G Networks: An Overview and Future Trends

Contact Info

Product

Resources

About