Unsupervised mmWave Beamforming via Autoencoders

Peken, Ture; Tandon, Ravi; Bose, T.

doi:10.1109/icc40277.2020.9149222

Cited by 10 publications

(6 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Achieves 30-70% gains in rates over supervised approaches. [182] • Compares digital beamforming methods in millimeter-wave environments.…”

Section: Channel Modeling and Analysismentioning

confidence: 99%

A Tutorial on NYUSIM: Sub-Terahertz and Millimeter-Wave Channel Simulator for 5G, 6G, and Beyond

Poddar,

Ju,

Shakya

et al. 2024

IEEE Commun. Surv. Tutorials

View full text Add to dashboard Cite

With the advancement of wireless communication to sub-terahertz (THz) and millimeter-wave (mmWave) bands, accurate channel models and simulation tools are becoming increasingly important for modeling a wide range of frequencies and scenarios. This paper provides a comprehensive tutorial on generating drop-based and spatial consistency-based channels using the open-source MATLAB-based NYU Channel Model Simulator (NYUSIM). NYUSIM is built on extensive real-world radio propagation measurements for the frequency range of 0.5-150 GHz, covering a variety of scenarios such as Urban Microcell (UMi), Urban Macrocell (UMa), Rural Macrocell (RMa), Indoor Hotspot (InH), and Indoor Factory (InF). Additionally, an overview of the evolution of simulators used to design and analyze wireless systems since the early days of cellular communication is also provided. We introduce the most popular types of simulators used in academia and industry, such as Channel Simulators (CSs), Link Level Simulators (LLSs), System Level Simulators (SLSs), and Network Simulators (NSs), to study wireless communication systems for 5G and beyond. Owing to the widespread adoption of the 3rd Generation Partnership Project (3GPP) Stochastic Channel Model (SCM) for channel generation in various simulators, we conduct a comparative analysis between the 3GPP SCM and NYUSIM channel model to highlight their differences. Moreover, NYUSIM's versatility extends beyond its MATLAB implementation, as it can be implemented in various LLSs, SLSs, and NSs, enabling researchers to incorporate realworld measurement-based channels into their simulations. To illustrate this capability, we showcase NYUSIM's implementation in ns-3, a widely used open-source discrete event network simulator. Additionally, we provide several applications of NYUSIM to highlight its potential uses.

show abstract

“…• Achieves 30-70% gains in rates over supervised approaches. [182] • Compares digital beamforming methods in millimeter-wave environments.…”

Section: Channel Modeling and Analysismentioning

confidence: 99%

A Tutorial on NYUSIM: Sub-Terahertz and Millimeter-Wave Channel Simulator for 5G, 6G, and Beyond

Poddar,

Ju,

Shakya

et al. 2024

IEEE Commun. Surv. Tutorials

View full text Add to dashboard Cite

show abstract

“…To countermeasure the aforementioned problem, unsupervised learning approaches were employed in several works (Kao et al, 2018;Peken et al, 2020b;Lin et al, 2020). In more detail, in (Peken et al, 2020b), an autoencoding-based SVD methodology was used in order to estimating the optimal beamforming codes at the TX and RX, while, in (Lin et al, 2020), Lin et al introduced a deep NN architecture for beamforming design that outperforms several previously presented deep learning approaches. Additionally, in (Kao et al, 2018), a MLbased clustering strategy with feature selection was employed to design three dimensional (3D) beamforming.…”

Section: Mac and Rrm Layermentioning

confidence: 99%

Machine Learning: A Catalyst for THz Wireless Networks

Boulogeorgos

Yaqub²,

Renzo

et al. 2021

Front. Comms. Net.

View full text Add to dashboard Cite

With the vision to transform the current wireless network into a cyber-physical intelligent platform capable of supporting bandwidth-hungry and latency-constrained applications, both academia and industry turned their attention to the development of artificial intelligence (AI) enabled terahertz (THz) wireless networks. In this article, we list the applications of THz wireless systems in the beyond fifth generation era and discuss their enabling technologies and fundamental challenges that can be formulated as AI problems. These problems are related to physical, medium/multiple access control, radio resource management, network and transport layer. For each of them, we report the AI approaches, which have been recognized as possible solutions in the technical literature, emphasizing their principles and limitations. Finally, we provide an insightful discussion concerning research gaps and possible future directions.

show abstract

“…Recently, the successful application of deep learning in various fields, particularly in computer science, has gained major attention in the communication community, and has promoted an increasing interest in applying it to address communication and signal processing problems [15]- [18]. The deep learning based intelligent communication paradigm has attained manifold accomplishments, including channel coding [19], random access [20], beamforming design [21]- [25] activity and signal detection [26], [27], autoencoder-based endto-end communication system [28], CSI feedback [29]- [31], and channel estimation [7], [32], [33], etc. To be specific, pure data-driven deep learning based solutions often employ deep neural networks (DNNs), including fully-connected neural networks and/or convolutional neural networks (CNNs), as a black box to design communication signal processing modules without any a priori model information.…”

Section: B Motivationsmentioning

confidence: 99%

“…By considering the multiuser channel estimation and feedback problem as a distributed source coding problem, the authors of [24] proposed a joint design of pilots and a novel DNN architecture, which mapped the feedback bits from all the users directly into the precoding matrix at the BS. By exploiting an unsupervised maching learning (ML) model, i.e., an autoencoder, the authors of [25] presented a linear autoencoder-based beamformer and combiner design, which maximizes the achievable rates over a mmWave channel. Moreover, in order to address the overwhelming feedback overhead of FDD massive MIMO systems, the authors of [29] proposed a CS-ReNet framework, where the CSI was first compressed at the users based on CS methods and then reconstructed at the BS using a deep learningbased recovery solver.…”

Section: B Motivationsmentioning

confidence: 99%

See 1 more Smart Citation

Model-Driven Deep Learning Based Channel Estimation and Feedback for Millimeter-Wave Massive Hybrid MIMO Systems

Ma,

Gao,

Gao

et al. 2021

Preprint

View full text Add to dashboard Cite

This paper proposes a model-driven deep learning (MDDL)-based channel estimation and feedback scheme for wideband millimeter-wave (mmWave) massive hybrid multipleinput multiple-output (MIMO) systems, where the angle-delay domain channels' sparsity is exploited for reducing the overhead. First, we consider the uplink channel estimation for time-division duplexing systems. To reduce the uplink pilot overhead for estimating high-dimensional channels from a limited number of radio frequency (RF) chains at the base station (BS), we propose to jointly train the phase shift network and the channel estimator as an auto-encoder. Particularly, by exploiting the channels' structured sparsity from an a priori model and learning the integrated trainable parameters from the data samples, the proposed multiple-measurement-vectors learned approximate message passing (MMV-LAMP) network with the devised redundant dictionary can jointly recover multiple subcarriers' channels with significantly enhanced performance. Moreover, we consider the downlink channel estimation and feedback for frequencydivision duplexing systems. Similarly, the pilots at the BS and channel estimator at the users can be jointly trained as an encoder and a decoder, respectively. Besides, to further reduce the channel feedback overhead, only the received pilots on part of the subcarriers are fed back to the BS, which can exploit the MMV-LAMP network to reconstruct the spatial-frequency channel matrix. Numerical results show that the proposed MDDL-based channel estimation and feedback scheme outperforms the stateof-the-art approaches.

show abstract

Unsupervised mmWave Beamforming via Autoencoders

Cited by 10 publications

References 16 publications

A Tutorial on NYUSIM: Sub-Terahertz and Millimeter-Wave Channel Simulator for 5G, 6G, and Beyond

A Tutorial on NYUSIM: Sub-Terahertz and Millimeter-Wave Channel Simulator for 5G, 6G, and Beyond

Machine Learning: A Catalyst for THz Wireless Networks

Model-Driven Deep Learning Based Channel Estimation and Feedback for Millimeter-Wave Massive Hybrid MIMO Systems

Contact Info

Product

Resources

About