Deep Learning for Channel Sensing and Hybrid Precoding in TDD Massive MIMO OFDM Systems

Attiah, Kareem M.; Sohrabi, Foad; Yu, Wei

doi:10.1109/twc.2022.3187790

Cited by 24 publications

(11 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A DNN-based approach for channel sensing and downlink hybrid beamforming is proposed in [14], which is generalized for any number of users. The multi-user cascaded CE is formulated as a denoising process in [15].…”

Section: A Deep Learning In Cementioning

confidence: 99%

Learn to Adapt to New Environments From Past Experience and Few Pilot Blocks

Wang

Gao

2023

IEEE Trans. Cogn. Commun. Netw.

View full text Add to dashboard Cite

In recent years, deep learning has been widely applied in communications and achieved remarkable performance improvement. Most of the existing works are based on data-driven deep learning, which requires a significant amount of training data for the communication model to adapt to new environments and results in huge computing resources for collecting data and retraining the model. In this paper, we will significantly reduce the required amount of training data for new environments by leveraging the learning experience from the known environments. Therefore, we introduce few-shot learning to enable the communication model to generalize to new environments, which is realized by an attention-based method. With the attention network embedded into the deep learningbased communication model, environments with different power delay profiles can be learnt together in the training process, which is called the learning experience. By exploiting the learning experience, the communication model only requires few pilot blocks to perform well in the new environment. Through an example of deep-learning-based channel estimation, we demonstrate that this novel design method achieves better performance than the existing data-driven approach designed for few-shot learning.

show abstract

Section: A Deep Learning In Cementioning

confidence: 99%

Learn to Adapt to New Environments From Past Experience and Few Pilot Blocks

Wang

Gao

2023

IEEE Trans. Cogn. Commun. Netw.

View full text Add to dashboard Cite

show abstract

“…We mention here that the beam alignment problem has also been investigated from the deep learning perspective in [22]- [26], but these works are all limited to the one-sided beam alignment problem. We remark that this paper considers a beam alignment problem within a coherence block where the Tx and Rx are not moving and the channel between the Tx and the Rx remains constant, while [27]- [29] address the beam tracking problem in the context of a time-varying channel model where the Rx moves following some Markov chain.…”

Section: A Related Workmentioning

confidence: 99%

“…DNN-based design with random sensing vectors [8], [26]: In this approach, we train two fully connected neural networks to directly map the received pilots to the optimized beamformers of the Tx and the Rx, respectively, in a centralized fashion. and [1024, 1024, 2M r ], respectively.…”

Section: B Benchmarksmentioning

confidence: 99%

Active Sensing for Two-Sided Beam Alignment and Reflection Design Using Ping-Pong Pilots

Jiang

Sohrabi

2023

IEEE J. Sel. Areas Inf. Theory

View full text Add to dashboard Cite

Beam alignment is an important task for millimeterwave (mmWave) communication, because constructing aligned narrow beams both at the transmitter (Tx) and the receiver (Rx) is crucial in terms of compensating the significant path loss in very high-frequency bands. However, beam alignment is also a highly nontrivial task because large antenna arrays typically have a limited number of radio-frequency chains, allowing only low-dimensional measurements of the high-dimensional channel. This paper considers a two-sided beam alignment problem based on an alternating ping-pong pilot scheme between Tx and Rx over multiple rounds without explicit feedback. We propose a deep active sensing framework in which two long short-term memory (LSTM) based neural networks are employed to learn the adaptive sensing strategies (i.e., measurement vectors) and to produce the final aligned beamformers at both sides. In the proposed ping-pong protocol, the Tx and the Rx alternately send pilots so that both sides can leverage local observations to sequentially design their respective sensing and data transmission beamformers. The proposed strategy can be extended to scenarios with a reconfigurable intelligent surface (RIS) for designing, in addition, the reflection coefficients at the RIS for both sensing and communications. Numerical experiments demonstrate significant and interpretable performance improvement. The proposed strategy works well even for the challenging multipath channel environments.

show abstract

“…Furthermore, even with CSI, finding the optimal EGT and equal ratio combining (EGC) beams requires a grid search over possible weights in the multiple-input multiple-output (MIMO) setting and is computationally prohibitive [19]. Recent works have explored directly predicting the hybrid BF weights for the BS without full CSI by learning sensing matrices [20] or a sequence of interactive sensing vectors based on feedback of previous measurements [21].…”

Section: B Site-specific Adaptationmentioning

confidence: 99%

“…The set of Tx and Rx probing beams are denoted by F and W. The first term U BF focuses on the BF gain of the synthesized beams and is the end-to-end objective. An obvious choice for the utility function is the average SNR or achievable rate of the predicted beam pairs, such as adopted in [20]. However, it tends to emphasize UEs with good channels more and neglect cell edge UEs.…”

Section: A Problem Formulationmentioning

confidence: 99%

Grid-Free MIMO Beam Alignment through Site-Specific Deep Learning

Heng¹,

Andrews²

2022

Preprint

View full text Add to dashboard Cite

Beam alignment is a critical bottleneck in millimeter wave (mmWave) communication. An ideal beam alignment technique should achieve high beamforming (BF) gain with low latency, scale well to systems with higher carrier frequencies, larger antenna arrays and multiple user equipments (UEs), and not require hard-to-obtain context information (CI). These qualities are collectively lacking in existing methods. We depart from the conventional codebook-based (CB) approach where the optimal beam is chosen from quantized codebooks and instead propose a grid-free (GF) beam alignment method that directly synthesizes the transmit (Tx) and receive (Rx) beams from the continuous search space using measurements from a few site-specific probing beams that are found via a deep learning (DL) pipeline.In realistic settings, the proposed method achieves a far superior signal-to-noise ratio (SNR)-latency trade-off compared to the CB baselines: it aligns near-optimal beams 100x faster or equivalently finds beams with 10-15 dB better average SNR in the same number of searches, relative to an exhaustive search over a conventional codebook.

show abstract

Deep Learning for Channel Sensing and Hybrid Precoding in TDD Massive MIMO OFDM Systems

Cited by 24 publications

References 38 publications

Learn to Adapt to New Environments From Past Experience and Few Pilot Blocks

Learn to Adapt to New Environments From Past Experience and Few Pilot Blocks

Active Sensing for Two-Sided Beam Alignment and Reflection Design Using Ping-Pong Pilots

Grid-Free MIMO Beam Alignment through Site-Specific Deep Learning

Contact Info

Product

Resources

About