Joint Deep Reinforcement Learning and Unfolding: Beam Selection and Precoding for mmWave Multiuser MIMO With Lens Arrays
Open accessThe DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index.
Abstract
The millimeter wave (mmWave) multiuser multiple-input multiple-output (MU-MIMO) systems with discrete lens arrays (DLA) have received great attention due to their simple hardware implementation and excellent performance. In this work, we investigate the joint design of beam selection and digital precoding matrices for mmWave MU-MIMO systems with DLA to maximize the sum-rate subject to the transmit power constraint and the constraints of the selection matrix structure. The investigated non-convex problem with discrete variables and coupled constraints is challenging to solve and an efficient framework of joint neural network (NN) design is proposed to tackle it. Specifically, the proposed framework consists of a deep reinforcement learning (DRL)-based NN and a deep-unfolding NN, which are employed to optimize the beam selection and digital precoding matrices, respectively. As for the DRL-based NN, we formulate the beam selection problem as a Markov decision process and a double deep Q-network algorithm is developed to solve it. The base station is considered to be an agent, where the state, action, and reward function are carefully designed. Regarding the design of the digital precoding matrix, we develop an iterative weighted minimum mean-square error algorithm induced deep-unfolding NN, which unfolds this algorithm into a layer-wise structure with introduced trainable parameters. Simulation results verify that this jointly trained NN remarkably outperforms the existing itera…
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“To reduce the model and sample complexity, as well as improve the performance and interpretability, unfolded neural networks were developed in [267]- [269] to parameterize the iterative policy via unfolding one iteration of the existing structured algorithm into one layer of a neural network.”
From Edge artificial intelligence for 6g: vision, enabling technologies, and applications (2022) · cites this paper as [268]
Citations
Cites
Human-level control through deep reinforcement learning
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. (2015) · Nature · doi:10.1038/nature14236 · Cited by 30,357 · cited 1×
“In addition, deep reinforcement learning (DRL) [27] has also been widely applied to solve the high-dimensional non-convex optimization problems.”
MethodNoncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas
Marzetta, T. L. (2010) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2010.092810.091092 · Cited by 6,302 · cited 1×
“3087233 of antennas at the transceiver in a cellular network, where the massive multiple-input multiple-output (MIMO) techniques can be readily realized to perform highly directional transmissions [2]–[4].”
BackgroundScaling Up MIMO: Opportunities and Challenges with Very Large Arrays
Rusek, F., Persson, D., Lau, B. K., Larsson, E. G., Marzetta, T. L., & Tufvesson, F. (2013) · IEEE Signal Processing Magazine · doi:10.1109/msp.2011.2178495 · Cited by 4,654 · cited 1×
“3087233 of antennas at the transceiver in a cellular network, where the massive multiple-input multiple-output (MIMO) techniques can be readily realized to perform highly directional transmissions [2]–[4].”
BackgroundScaling up MIMO: Opportunities and challenges with very large arrays
Rusek, F., Persson, D., Lau, B. K., Larsson, E. G., Marzetta, T. L., Edfors, O., Tufvesson, F., Print, P., Rusek, F., Persson, D., et al. (2013) · Cited by 4,089
Five disruptive technology directions for 5G
Boccardi, F., Heath, R. W., Lozano, A., Marzetta, T. L., & Popovski, P. (2014) · IEEE Communications Magazine · doi:10.1109/mcom.2014.6736746 · Cited by 3,709
Spatially Sparse Precoding in Millimeter Wave MIMO Systems
Ayach, O. E., Rajagopal, S., Abu-Surra, S., Pi, Z., & Heath, R. W. (2014) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2014.011714.130846 · Cited by 3,557 · cited 1×
“The hybrid analog-digital (AD) transceiver consisting of cascaded baseband digital precoder and RF analog beamformer has been proposed as a trade-off between cost and performance, which employs less RF chains than the number of antennas, while imposing a constant modulus constraint on the analog beamforming matrix [5].”
BackgroundDeep Reinforcement Learning with Double Q-Learning
Van Hasselt, H., Guez, A., & Silver, D. (2016) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v30i1.10295 · Cited by 3,318 · cited 2×
“We apply the DDQN [36] to learn the target value y t to avoid the overestimation and make the NN stable, which is expressed as where θ (cid:48) and θ denote the parameters of the two NNs, which have been applied to select and evaluate the action, respectively.”
MethodShi, Q., Razaviyayn, M., Luo, Z.-Q., & He, C. (2011) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2011.2147784 · Cited by 2,129 · cited 5×
“Given the beam selection matrix, the problem of the digital precoding design is a classic sum-rate maximization problem with power constraints that can be settled by the iterative WMMSE algorithm [36].”
BackgroundMethodInfluentialHuang, L., Bi, S., & Zhang, Y.-J. A. (2020) · IEEE Transactions on Mobile Computing · doi:10.1109/tmc.2019.2928811 · Cited by 1,040 · cited 5×
“The DQN has been widely applied in mobile offloading [28], dynamic channel access [29], radio access [30], resource allocation [31], [32], and mobile-edge computing (MEC) [33].”
BackgroundMethodInfluentialLearning to Optimize: Training Deep Neural Networks for Interference Management
Sun, H., Chen, X., Shi, Q., Hong, M., Fu, X., & Sidiropoulos, N. D. (2018) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2018.2866382 · Cited by 874 · cited 2×
“Recently, the machine learning techniques emerge as an efficient method and there are a number of studies employing the neural networks (NNs) for solving the investigated problems [16], such as precoding design [17], beam selection [18]–[20], and resource allocation [21].”
BackgroundMethod
Cited by
Edge artificial intelligence for 6g: vision, enabling technologies, and applications
Letaief, K. B., Shi, Y., Lu, J., & Lu, J. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126076 · Cited by 829
“To reduce the model and sample complexity, as well as improve the performance and interpretability, unfolded neural networks were developed in [267], [268], [269] to parameterize the iterative policy via unfolding one iteration of the existing structured algorithm into one layer of a neural network.”
MethodA Survey of Beam Management for mmWave and THz Communications Towards 6G
Xue, Q., Ji, C., Ma, S., Guo, J., Xu, Y., Chen, Q., & Zhang, W. (2024) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2024.3361991 · Cited by 280
“In [88], the joint design of beam selection and digital precoding for mmWave MU-MIMO systems was investigated to maximize the sum-rate.”
MethodLearning Decentralized Wireless Resource Allocations With Graph Neural Networks
Wang, Z., Eisen, M., & Ribeiro, A. (2022) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2022.3163626 · Cited by 100
“And, if the parametrization is suitably chosen, they can scale to networks with large numbers of transceivers [17]–[27].”
Sun, R., Cheng, N., Li, C., Quan, W., Zhou, H., Wang, Y., Zhang, W., & Shen, X. (2025) · IEEE Communications Surveys & Tutorials · doi:10.1109/comst.2025.3574765 · Cited by 93
“To reduce online processing time and keep interpretability, a general framework of iterative WMMSE-unfolded customized DNN is developed in matrix form for the multi-user MIMO scenario [104], [105], [116].”
Deep Unfolding Hybrid Beamforming Designs for THz Massive MIMO Systems
Nguyen, N. T., Ma, M., Lavi, O., Shlezinger, N., Eldar, Y. C., Swindlehurst, A. L., & Juntti, M. J. (2023) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2023.3322852 · Cited by 79
“The former relies mainly on the learning capability of deep neural networks (DNNs) [24], [25], [26], convolutional neural networks (CNNs) [27], [28], [29], [30], [37], [38], [39], [40], or deep reinforcement learning [41], [42] to generate HBF beamformers.”
BackgroundData and domain knowledge dual‐driven artificial intelligence: Survey, applications, and challenges
Nie, J., Jiang, J., Li, Y., Wang, H., Erċışlı, S., & Lv, L. (2023) · Expert Systems · doi:10.1111/exsy.13425 · Cited by 76
“Hu et al. developed an efficient framework for joint neural network (NN) design (Hu et al., 2021), consisting of a deep reinforcement learning (DRL)-based NN (data-driven) and a deep unfolding NN (knowledge mining) for pre-training and fitting in order to reduce the computational complexity of…”
Radar-Assisted Predictive Beamforming for UAV-Aided Networks: A Deep-Learning Solution
Yin, R., Peng, J., Cai, Y., Wu, C., Champagne, B., & Al‐Dhahir, N. (2025) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2025.3572037 · Cited by 56
“Inthepastfewyears,thedeep-unfoldingnetwork(DUN) hasemergedasaneffectiveapproachtosolvevarioustasksin differentcommunicationscenarios,suchasmulti-inputmulti-output(MIMO)detection[22],powercontrol[19],[23],[24], [25]andbeamforming[26],[28].”
Self-citationNarottama, B., & Shin, S. Y. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3280459 · Cited by 52
“More often than not, these kinds of setbacks can be mitigated through feature preprocessing techniques, in particular, reduction of the input dimensions [29], [30].”
BackgroundLearn to Rapidly and Robustly Optimize Hybrid Precoding
Lavi, O., & Shlezinger, N. (2023) · IEEE Transactions on Communications · doi:10.1109/tcomm.2023.3292472 · Cited by 52
“DNN-aided optimization of hybrid precoders was considered in [21], [22], [23], [24], [25], [26], [27], [28], [29], and [30], which employed multi-layered perceptrons [21], [22] and convolutional neural networks [23], [24], [25], [26], [27], [28] as the optimizer, while [29], [30] employed deep…”
MethodDeep-Unfolding Beamforming for Intelligent Reflecting Surface Assisted Full-Duplex Systems
Liu, Y., Hu, Q., Cai, Y., Yu, G., & Li, G. Y. (2021) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2021.3133296 · Cited by 50
“Hence, the deep-unfolding has attracted great research interests and has a wide range of applications in communications, such as signal detection [37]– [39], resource allocation [40], [41], and precoding [42]–[45].”
BackgroundSelf-citationTwo-Timescale End-to-End Learning for Channel Acquisition and Hybrid Precoding
Hu, Q., Cai, Y., Kang, K., Yu, G., Hoydis, J., & Eldar, Y. C. (2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3126050 · Cited by 41
Self-citationCai, C., Yuan, X., & Zhang, Y. A. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3531575 · Cited by 39
“Building upon this, we propose a deep unfolding network for precoding design, which primarily follows the methodologies discussed in [30], [31], and [32].”
Hou, Q., Cai, Y., Hu, Q., Lee, M., & Yu, G. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3217176 · Cited by 37
“Inspired by the training strategy in [22], we jointly train the proposed two NNs by the following two steps: (i) The reward function of the APC network is computed based on P and q , which are the outputs of the deep-unfolding NN. (ii) One of the inputs of the deep-unfolding NN, A , is obtained by…”
BackgroundMethodSelf-citationMixed-Timescale Deep-Unfolding for Joint Channel Estimation and Hybrid Beamforming
Kang, K., Hu, Q., Cai, Y., Yu, G., Hoydis, J., & Eldar, Y. C. (2022) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2022.3191124 · Cited by 34
Self-citationComprehensive Review of Deep Unfolding Techniques for Next-Generation Wireless Communication Systems
Deka, S., Deka, K., Nguyen, N. T., Sharma, S., Bhatia, V., & Rajatheva, N. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3652234 · Cited by 32
“To address this, Hu et al. [176] proposed a joint design approach combining DL and model-driven techniques.”
DDPG-Driven Deep-Unfolding With Adaptive Depth for Channel Estimation With Sparse Bayesian Learning
Hu, Q., Shi, S., Cai, Y., & Yu, G. (2022) · IEEE Transactions on Signal Processing · doi:10.1109/tsp.2022.3207269 · Cited by 32
“Moreover, to solve the mixed integer non-linear problems, deep Q-network (DQN) [13] has been widely employed in communications, such as beam selection [14], dynamic channel access [15], and resource allocation [16], where the problems are modeled as Markov decision process (MDP).”
MethodSelf-citationWang, L., Yang, F., Chen, Y., Lai, S., & Wu, W. (2023) · Physical Communication · doi:10.1016/j.phycom.2023.102013 · Cited by 32
Jie, H., Yang, C., Yang, F., Zhang, S., Tolba, A. K. R., Jolfaei, A., & Yu, K. (2024) · Digital Communications and Networks · doi:10.1016/j.dcan.2024.10.003 · Cited by 28
AI Empowered Resource Management for Future Wireless Networks
Shen, Y., Zhang, J., Song, S. H., & Letaief, K. B. (2021) · 2021 IEEE International Mediterranean Conference on Communications and Networking (MeditCom) · doi:10.1109/meditcom49071.2021.9647580 · Cited by 22
Deep Reinforcement Learning for Multi-User Massive MIMO With Channel Aging
Feng, Z., & Clerckx, B. (2023) · IEEE Transactions on Machine Learning in Communications and Networking · doi:10.1109/tmlcn.2023.3325299 · Cited by 22
“In terms of active beamforming using DRL, several efforts have been made on designing low complexity algorithms based on deep Q-network (DQN) [28], [34]–[36] and partially observed MDP [38] frameworks.”
MethodPiecewise-DRL: Joint Beamforming Optimization for RIS-Assisted MU-MISO Communication System
Li, J., Wang, W., Jiang, R., Wang, X., Fei, Z., Huang, S., & Li, X. (2023) · IEEE Internet of Things Journal · doi:10.1109/jiot.2023.3275818 · Cited by 20
“By employing a DRL-NN framework, Hu et al. [41] jointly designed beam selection and precoding matrix to maximize the system sum-rate in an MIMO communication system.”
MethodYe, N., Cao, X., Ding, X., Li, J., Zhao, D., & Ouyang, Q. (2023) · IEEE Transactions on Green Communications and Networking · doi:10.1109/tgcn.2023.3299567 · Cited by 19
“…spectrum is also limited [24], most of the studies on wireless communication systems with lens antenna array focus on optimizing spectrum-efficiency (SE) or sum rate [20], [21], [22], [23] rather than EE, which leads to inefficiency of these schemes for energy-constrained transmission systems.”
BackgroundFast Beampattern Synthesis Algorithm for Flexible Conformal Array
Tian, X., Chen, H., He, M., & Wang, W.-Q. (2022) · IEEE Signal Processing Letters · doi:10.1109/lsp.2022.3223804 · Cited by 18
“More novelly, [13] addressed NP-hard and non-convex beam problem with a deep reinforcement learning (DRL)-based neural network (NN) and a deep-unfolding NN.”
BackgroundZhang, J., Xu, S., Li, C., Huang, Y., & Yang, L. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3650204 · Cited by 17
“To be specific, authors in [22] developed a DRL-based beam training method, interacting with the environment to capture dynamic spatial patterns and decide the beam search range without precise CSI.”
Multi-UE Multi-AP Beam Alignment in User-Centric Cell-Free Massive MIMO Systems Operating at mmWave
Buzzi, S., D’Andrea, C., Fresia, M., & Wu, X. (2022) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2022.3170787 · Cited by 17
“Papers [29]–[31] tackle the problem of BA and beam selection in a multiuser scenario applying machine learning tools to solve the problem.”
BackgroundHybrid Beamforming Design for Covert mmWave MIMO With Finite-Resolution DACs
Ci, W., Qi, C., & You, X. (2025) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2025.3531523 · Cited by 13
“Fortunately, millimeter wave (mmWave) communications inherently provide the covertness due to the sparse scattering characteristics [2], [3].”
Cui, X., Zhang, Z., Li, J., Jiang, B., Li, S., & Liu, J. (2023) · Physical Communication · doi:10.1016/j.phycom.2023.102207 · Cited by 12
Yan, Y., Zhang, B., Li, C., Bai, J., & Yao, Z. (2023) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2023.3280910 · Cited by 11
“Single-agent based RL algorithms [19], [20], [21], [22], [23], [24] have been proposed to address the problem of optimal hybrid beamforming in massive MIMO mmWave networks.”
BackgroundMethodHe, B., Zhang, A., Hao, W., Liu, P., & Ning, B. (2022) · IEEE Transactions on Circuits & Systems II Express Briefs · doi:10.1109/tcsii.2022.3213855 · Cited by 10
“Additionally, there have been several works studying the beam selection strategies under lens antenna array structure, including single beam selection [10], [11], and multiple beam selection via multiple RF (MBMRF) chains [13], [14].”
BackgroundHuang, Y., Chi, K., Yang, Q., Yang, Z., & Zhang, Z. (2024) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2024.3464509 · Cited by 9
“Hu et al. [9] developed a hybrid precoder framework, where beam selection was performed by a DRL network, and the digital precoding part was implemented by a deep folding network of the weighted minimum mean square error (WMMSE) algorithm.”
DeepFP: Deep-Unfolded Fractional Programming for MIMO Beamforming
Zhu, J., Chang, T., Xiang, L., & Shen, K. (2026) · IEEE Transactions on Communications · doi:10.1109/tcomm.2026.3653777 · Cited by 7
“Since two methods are both iteratively structured, a natural idea is to learn their behaviors via deep unfolding, as pursued extensively in [6], [7], [8], [9], [10], [11], [12], [13].”
InfluentialLearning-Based Joint Channel Prediction and Antenna Selection for Massive MIMO with Partial CSI
He, K., Vu, T. X., Chatzinotas, S., & Ottersten, B. (2022) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps56602.2022.10008768 · Cited by 7
“Solutions to problem (3) have been well studied in the literature [4], [6], [7], [11]–[13] under the complete CSI assumption, meaning that the wireless channel is fully observable.”
BackgroundLopukhova, E. A., Abdulnagimov, A., Voronkov, G. S., Kutluyarov, R. V., & Grakhova, E. P. (2023) · Information · doi:10.3390/info14020086 · Cited by 6
Deep Contextual Bandits Learning-Based Beam Selection for mmWave MIMO Systems
Mohsenivatani, M., Ali, S., Rajatheva, N., & Latva‐aho, M. (2023) · IEEE International Mediterranean Conference on Communications and Networking (MeditCom) · doi:10.1109/meditcom58224.2023.10266618 · Cited by 6
“state in [8] to model the problem as MDP.”
BackgroundLu, Y., Ou, K.-H., Chen, H.-Y., Wu, M.-H., Yang, T.-W., Tsai, H.-H., & Chou, C. (2022) · IEEE Vehicular Technology Conference · doi:10.1109/vtc2022-spring54318.2022.9860620 · Cited by 6
Reinforcement Learning With Selective Exploration for Interference Management in mmWave Networks
Dinh‐Van, S., Nguyen, V.-L., Bulut, B., Masaracchia, A., & Higgins, M. D. (2025) · IEEE Transactions on Machine Learning in Communications and Networking · doi:10.1109/tmlcn.2025.3537967 · Cited by 5
“…(RL) offers a fundamentally different approach to problem-solving compared to conventional methods, and it has been successfully applied to various problems at physical (PHY) layer such as beamforming, power control, and interference management in diverse settings [18], [19], [20], [21], [22].”
Low-Complexity Beam Selection Algorithms Based on SVD for mmWave Massive MIMO Systems
Yu, J., Yang, J., Wang, S., Cai, Y., & Liu, J. (2023) · IEEE Communications Letters · doi:10.1109/lcomm.2023.3292126 · Cited by 5
“In fact, it is a mixed-integer non-linear programming (MINLP) problem since it involves a discrete variable F [10].”
BackgroundSelf-citationSecrecy Rate Maximization for Active Reconfigurable Intelligent Surface Assisted MIMO Systems
Gao, B., Zhao, J., Yan, S., & Xiao, S. (2024) · IEEE Access · doi:10.1109/access.2024.3416553 · Cited by 5
“Hu et al. [27] investigated the joint optimization of BS precoder and beam selection via a deep Q-network (DQN) approach.”
MethodXu, R., Lin, F., Shao, W., Wang, H., Meng, F., & Li, J. (2024) · Entropy · doi:10.3390/e26040331 · Cited by 5
“This approach allows the trained algorithm to consider the seasonal variations in renewable energy generation and electricity demand.”
BackgroundDeep-Unfolding for Next-Generation TransceiversPreprint
Hu, Q., Cai, Y., Zhang, G., Yu, G., & Li, G. Y. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.08303 · Cited by 5
“Particularly, inspired by the first-order Taylor expansion of the inverse matrix, A−1 can be approximated by A†X + AY + Z, where X, Y, and Z are introduced trainable parameters [6].”
BackgroundMethodInfluentialSelf-citationYang, R., Li, X., Huang, Y., Yang, L., & Zhang, W. (2025) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2025.3592854 · Cited by 4
“Deep reinforcement learning (DRL) has also gained increasing attention for the ability to solve non-convex problem [12], [31], [32].”
Ramanathan, S., & Bennet, M. A. (2024) · Telecommunication Systems · doi:10.1007/s11235-024-01135-4 · Cited by 4
“We concentrate on increasing the system’s downlink sum rate in the downlink precoder design [24].”
Machine Learning-Enabled Joint Codebook Design and Beam Selection
Liang, F., & Cai, Y. (2024) · International Symposium on Wireless Communication Systems · doi:10.1109/iswcs61526.2024.10639069 · Cited by 4
Self-citationHashemi, R., Ranasinghe, V., Veijalainen, T., Kela, P., & Wichman, R. (2024) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc59610.2024.10817344 · Cited by 4
“A deep Q-network (DQN)-based method for choosing the candidate beam with real/imaginary parts of the channel matrix information as state space was discussed in [5].”
GNN-Enabled Deep Unfolding for Precoding in Massive MIMO LEO Satellite Communications
Zhou, H., Gong, X., Tsinos, C. G., You, L., Gao, X., & Ottersten, B. (2025) · IEEE Wireless Communications and Networking Conference · doi:10.1109/wcnc61545.2025.10978590 · Cited by 4
“Algorithm unfolding addresses these issues by linking classic iterative algorithms with NNs, converting algorithm parameters into trainable ones, improving layer interpretability, and offering theoretical guarantees for better reliability and predictability [13].”
Deep Learning for Velocity and Range Estimation of Drones in Urban Occlusion Environments
Xu, S., Yuan, W., & Cai, Y. (2026) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2026.3658147 · Cited by 3
“Hu et al. un-fold iterative transceiver/precoder solvers into trainable networks to accelerate real-time inference [36]; Hu et al. use DRL for beam selection and link adaptation in non-stationary mmWave/multiuser settings [37].”
Self-citationA New Low-Complexity WMMSE Algorithm for Downlink Massive MIMO Systems
Zhou, N., Wang, Z., He, L., & Huang, Y. (2022) · International Conference on Wireless Communications and Signal Processing (WCSP) · doi:10.1109/wcsp55476.2022.10039088 · Cited by 3
“Both [8] and [9] use first-order Taylor expansion with several trainable matrices to approximate the matrix inversion.”
MethodDeep Learning for MIMO Communications
Cai, Y., Hu, Q., Zhang, G., & Kang, K. (2023) · IntechOpen eBooks · doi:10.5772/intechopen.112038 · Cited by 3
Self-citationMeta Reinforcement Learning for Fast Spectrum Sharing in Vehicular NetworksPreprint
Huang, K., Liang, L., Jin, S., & Li, G. Y. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.17185 · Cited by 3
“Nowadays, RL has been widely leveraged to address many challenging problems in wireless communications, such as beam selection and precoding in MIMO systems [10], phase shift design in reconfigurable intelligent surface (RIS) assisted systems [11] and wireless resource allocation [12].”
Liang, F., Cai, Y., Liu, A., & Champagne, B. (2025) · International Conference on Communications · doi:10.1109/icc52391.2025.11161416 · Cited by 2
“Specifically, we adopt the WMMSE algorithm for short-term beamforming design [8], and then optimize the long-term APV using SSCA [9] with the updated short-term beamforming matrix.”
Self-citationTwo-Timescale Deep Optimization of Positioning and Beamforming in Movable Antenna Arrays
Liang, F., Cai, Y., Liu, A., & Champagne, B. (2026) · IEEE Transactions on Wireless Communications · doi:10.1109/twc.2026.3675048 · Cited by 2
“As a result, DUNNs have gained increasing attention in communication systems, with successful applications in resource allocation [34], [35], signal detection [36], [37], [38], and beamforming [27], [28], [39], [40].”
Self-citationHu, Q., Shi, S., Cai, Y., & Yu, G. (2022) · International Conference on Communications · doi:10.1109/icc45855.2022.9839060 · Cited by 2
“In particular, the iterative algorithms are unfolded into a layer-wise architecture similar to deep neural networks (DNNs) [9]–[11].”
MethodSelf-citationLi, Z., Zhou, X., Wan, C., Du, G., & Wang, Y. (2023) · Applied Sciences · doi:10.3390/app131910855 · Cited by 2
“In [34], the authors design the selection and precoding matrices jointly for millimeter-wave systems.”
BackgroundXu, X., Zheng, H., Du, M., Feng, X., & Chen, Y. (2025) · International Conference on Communications · doi:10.1109/icc52391.2025.11161660 · Cited by 2
“The beam selection problem is formulated as a MDP [7], where the MDP is defined as a quintuple ( S , A , R , P , γ ) , where S represents the state space, A denotes the action space, R : S → R indicates the reward function, P : S ×A×S → R signifies the state transition probability, and γ ∈ (0 , 1)…”
Xiao, B., Hu, S., Chen, X., Zhai, Z., Li, B., Ni, W., Wang, X. E., & Hossain, E. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2606.24424 · Cited by 2
“…(drQ) [83]; DSP [88]; PIRL [91] Beam Selection / Tracking (RL and bandit) DQN/DDQN/PG beam selection [191]–[194]; joint beam selection + precoding [195]; bandit beam–rate tracking [196], [197] Integrated Gradients [198]; PIRL [91]; DSP [88] Interference Alignment (RL) DRL beamforming and power…”
Environment-Aware Transfer Reinforcement Learning for Sustainable Beam Selection
Salami, D., Hashemi, R., Kazemi, P., & Uusitalo, M. A. (2025) · IEEE Globecom Workshops (GC Wkshps) · doi:10.1109/gcwkshps68340.2025.11591027 · Cited by 1
“Reinforcement Learning (RL) is essential for beam selection in 5G and beyond due to its ability to handle the dynamic and complex nature of modern wireless environments [3].”
Machine Learning-Based Adaptive Codebook Design and Beamforming for Near-Field Communications
Zhang, M., Cai, Y., Yu, G., Xu, J., & Swindlehurst, A. L. (2025) · IEEE Transactions on Communications · doi:10.1109/tcomm.2025.3637068 · Cited by 1
“Besides, following the analysis in [39], the joint NN introduces a complexity of O ( NK + L ( K 2 M 2 + KM 2 .”
Self-citationIoannou, I. I., Georgiades, M., Nagaradjane, P., Khalifeh, A. F., Christophorou, C., Raspopoulos, M., & Vassiliou, V. (2026) · Network · doi:10.3390/network6020035 · Cited by 1
“Hu et al. [14] proposed a joint deep reinforcement learning and deep unfolding architecture for beam selection and precoding in millimeter-wave multiuser MIMO systems equipped with lens antenna arrays.”
Jagadhabi, N. (2026) · Journal of Wireless Mobile Networks Ubiquitous Computing and Dependable Applications · doi:10.58346/jowua.2026.i2.007 · Cited by 1
Zhang, W., Qian, Y., Xu, W., Ng, D. W. K., & You, X. (2023) · IEEE/CIC International Conference on Communications in China (ICCC) · doi:10.1109/iccc57788.2023.10233509 · Cited by 1
Deep Reinforcement Learning for Enhancing the Secrecy of a MU-MISO UOWC Network
Illi, E., Baccour, E., Qaraqe, M., & Hamdi, M. (2023) · Globecom · doi:10.1109/globecom54140.2023.10437117 · Cited by 1
“To this end, DRL has brought notable intuitions for using it in problems such as optimizing the beamforming and precoding matrices in multi-user (MU)- MIMO networks [11].”
MethodDeep Unfolding Based Precoding for Task-Oriented Semantic Communications Over MIMO Channels
Cai, C., Yuan, X., & Zhang, Y. A. (2024) · Asilomar Conference on Signals, Systems and Computers (ACSSC) · doi:10.1109/ieeeconf60004.2024.10942932 · Cited by 1
“To resolve this issue, we use PyTorch to manually implement the required complex operations by their corresponding real-valued representations [15].”
2021 Index IEEE Journal on Selected Areas in Communications Vol. 39
(2021) · IEEE Journal on Selected Areas in Communications · doi:10.1109/jsac.2021.3132125
Lopukhova, E. A., Kuznetsova, E. V., Voronkov, G. S., Kutluyarov, R. V., & Grakhova, E. P. (2023) · International Siberian workshops and tutorials on electron devices and materials · doi:10.1109/edm58354.2023.10225199
“To date, algorithms based on machine learning (ML) have demonstrated high efficiency in IoT applications [2], including in problems of location determination and improved positioning accuracy [20]–[22].”
MethodShu-peng, J., Li, X., Gao, N., & Jin, S. (2023) · IEEE International Conference on Communication Technology (ICCT) · doi:10.1109/icct59356.2023.10419517
“Combining numerical optimization algorithms with DL, physical layer solutions based on model-driven DL, such as deep unfolding (DU) [7], have the advantages of relatively low training time, reduced risk of overfitting, and rapid implementation.”
MethodXue, Q., Dong, C., Niu, K., Liang, Z., Li, X., Suo, S., & Gao, Q. (2024) · IEEE Transactions on Vehicular Technology · doi:10.1109/tvt.2024.3436555
“This scheme predefines multiple beam directions (i.e., using a codebook), and the beam is selected from the codebook for each user terminal [9], [10], [11].”
AI-Empowered Mode Selection and Beamforming for STAR-RIS-Assisted Communications
Zhang, M., Cai, Y., & Swindlehurst, A. L. (2024) · IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC) · doi:10.1109/pimrc59610.2024.10817397
“Specifically, the proposed joint NN design consists of an advantage pointer-critic (APC) NN [15], a fully connected NN, and a deep-unfolding NN [16].”
Self-citationEnergy Efficiency and Optimization of RIS‐Based Wireless Communication Systems
Ho, K., Nguyen, M. D., Pham, M. T., Truong, H. G., & Pogaku, A. C. (2025) · Wiley eBooks · doi:10.1002/9781394250141.ch17
Energy-Efficient Beamforming for Satellite MIMO Transmission
Li, J., Ding, X., & Ye, N. (2025) · Springer eBooks · doi:10.1007/978-981-96-5951-7_7
Distributed Network Control Using Machine Learning
Tehrani, P. (2023) · eScholarship (California Digital Library)
Baseband-Efficient WMMSE Precoding: From a Signal Weighting Cost PerspectivePreprint
Gao, S., Xu, F., Li, M., Ning, X., Qiu, L., Ning, B., & Shi, Q. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.18368
Bian, Y.-J., Guo, W., Yang, J., Song, S.-H., Zhang, J., Jin, S., & Letaief, K. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2606.26567 · Cited by 2
“To limit strong multi-user interference, the action space should be constrained against highly correlated beam assignments that cause equivalent channel rank degradation [8], [29].”
InfluentialAI-Empowered Fog or Edge Resource Management
Banchhode, S., Phadatare, A., & Varade, P. (2026) · Smart innovation, systems and technologies · doi:10.1007/978-3-032-15996-0_20
Metadata
- Authors
- Hu, Q., Liu, Y., Cai, Y., Yu, G., & Ding, Z.
- Affiliations
- Zhejiang University (CN)
- University of California, Davis (US)
- Venue
- IEEE Journal on Selected Areas in Communications
- Published
- 2021-08
- Volume
- 39
- Issue
- 8
- Pages
- 2289-2304
- ISSN
- 0733-8716 (ISSN-L)
- 1558-0008
- OpenAlex
- W3167193327
- Cited by
- 64
- References
- 41
- Citations by year
- 2026: 7 · 2025: 12 · 2024: 11 · 2023: 18 · 2022: 11 · 2021: 5
- FWCI
- 3.31
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Journal on Selected Areas in Communications (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (61971376, 61831004)
- Zhejiang Provincial Natural Science Foundation for Distinguished Young Scholars (LR19F010002)
- Topics
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Millimeter-Wave Propagation and Modeling
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced Wireless Communication Technologies
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Advanced MIMO Systems Optimization
- Keywords
- millimeter wave, multi-user MIMO, beam selection, digital precoding, deep reinforcement learning, deep unfolding, double deep Q-network, sum rate maximization, Markov decision process, neural network design
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.