The extreme multiple-input multiple-output (X-MIMO) framework is vital for 6G due to its capacity for extensive spatial multiplexing using numerous antenna ports. As antenna ports increase, the need for channel state information reference signals (CSI-RS) increases, leading to high overhead and energy usage, which limits data resources and performance gains. Although engaging all antennas ports for data transmission and reception reaps the high performance of X-MIMO, channel measurement from the reduced number of ports increases the efficiency. This study introduces a deep learning-based method for CSI reconstruction that reduces overhead in X-MIMO systems, called Transformer-based CSI feedback with subsampled CSI-RS (TS-CSI-RS). The TS-CSI-RS reconstructs the full-port channel information from undersampled measurement via Transformer. Simulations in a realistic channel environment show that TS-CSI-RS significantly reduces CSI-RS overhead (up to 75%) compared to conventional 5G systems, allowing spatial adaptation without activating all antenna ports for CSI measurement.

Modern base stations exploit distributed architectures defined by the O-RAN standard for efficient operation. In ORAN, the base station functionality is split between a radio unit (RU) and a distributed unit (DU). Further, based on 3GPP split option 7.2, O-RAN specifies two categories of RU: Cat-A (or 7.2A) wherein precoding is performed at the DU, and Cat-B (or 7.2B) wherein precoding operation is offloaded to the RU. Hence, in 7.2B, the precoding information is transferred from DU to RU, but, after application of beamspace compression (BSC). This involves conversion of the precoders to beamspace, which is followed by quantization of its real and imaginary coefficients. However, the BSC approach is suboptimal as it does not fully exploit the underlying correlation in the precoding vectors. To address this limitation, we propose a deep learning (DL) based beamspace precoder compression approach for O-RAN. Using 3GPP compliant simulations, we show that the proposed approach decreases the precoder transfer overhead by up to 85%, while also achieving a spectral efficiency gain of up to 25%. Furthermore, we show that by compressing the precoders in the beamspace, the operational complexity can be drastically reduced, specifically due to reduction in auto-encoder (AE) input dimensionality.

This work proposes a deep learning-based data-aided active user detection network (D-AUDN) for grant-free sparse code multiple access (SCMA) systems that leverages both SCMA codebook and Zadoff-Chu preamble for activity detection. Due to disparate data and preamble distribution as well as codebook collision, existing D-AUDNs experience performance degradation when multiple preambles are associated with each codebook. To address this, a user activity extraction network (UAEN) is integrated within the D-AUDN to extract a-priori activity information from the codebook, improving activity detection of the associated preambles. Additionally, efficient SCMA codebook design and preamble sequence association are considered to further enhance performance.

Deep learning (DL) has emerged as a transformative tool in wireless communications, offering powerful methods for optimizing complex systems. One critical challenge in deploying DL-based solutions is ensuring that models can generalize effectively across various channel conditions. This paper investigates the generalization capabilities of a DL-based Collision-Aware Multi-User Detection (CA-MUD) system within Grant-Free Sparse Code Multiple Access (GF-SCMA) frameworks, which are vital for efficient massive machine-type communication (mMTC). By exploring different training environments, such as independent and identically distributed (IID) Rayleigh and tapped delay line (TDL) channels, the study aims to identify optimal strategies to maintain robust communication performance under diverse conditions.

In grant-free sparse code multiple access (SCMA) systems, SCMA codebooks (CBs) are used for efficient grant-free random access. However, CB collisions can occur when multiple active users select the same CB, degrading the performance of multi-user detection (MUD) at the base station (BS). The existing methods modify the factor graph on the message-passing algorithm (MPA) for each CB collision scenario, resulting in high computational complexity. In this paper, we aim to confirm that even in the presence of CB collisions, MUD performance can be ensured through a deep learning (DL)-based receiver and explore its limitations. We propose a single DL architecture for collision-aware MUD (CA-MUD) that can tolerate CB collisions, without resorting to the distinct MUD processes associated with individual collision scenarios. To facilitate the generation of training data for CA-MUD that comprehensively represents the grant-free SCMA scenario, we introduce a transceiver model that regulates the number of active CBs and sets the maximum tolerable CB collisions. Simulation results demonstrate that our proposed approach allows a single CA-MUD network to handle various CB collision scenarios, including 2-fold CB collision subject to a limited number of active users.

We consider a deep learning (DL) based approach to optimize channel state information (CSI) reporting in massive multiple-input and multiple-output (MIMO) systems. For CSI compression, existing methods use frequency-domain (FD) and spatial-domain (SD) correlation, whereas correlation also exists in the time-domain (TD) under fading conditions. Hence, we propose a DL-based three-dimensional compression (DL-3DC) approach to improve CSI reporting accuracy by using correlation in FD, SD and TD. In addition, to reduce the feedback overhead, we propose two DL-based CSI prediction methods: eigenvector based and MIMO channel based. We then integrate CSI prediction with DL-3DC at the UE and propose a joint CSI compression and prediction (JCCP) scheme to improve the CSI accuracy and reporting overhead trade-off. Although, UE-side JCCP is efficient when the future CSI application time instance is known in advance. To support prediction when this is unknown, we propose BS-side JCCP to predict the CSI at the BS after it is received. Through simulations, we show that DL-3DC improves the CSI reporting accuracy by up to 13.8% compared to the latest CSI reporting mechanism in New Radio (NR). Further, we show that JCCP significantly reduces the reporting overhead by up to 87.5% compared to NR CSI reporting.

To tackle the power consumption challenges in terahertz band wireless communication, this study proposes a deep learning-driven approach for transceiver design that utilizes one-bit quantization and oversampling at the receiver. The solution also involves implementing Faster-than-Nyquist (FTN) transmission on a fading channel. Our approach employs a convolutional autoencoder (AE) to enable the transmission of higher-order modulation over a one-bit fading channel while utilizing pilots. By exploiting the AE transceiver, it is evident that performance in quantized communication has significantly improved for QPSK, 16-QAM, and 64-QAM modulation levels, approaching the theoretical lower bound for the corresponding modulation over additive white Gaussian noise (AWGN) channel. Furthermore, the study has explored how to use the robust error-correcting capabilities of the AE transceiver to boost spectral efficiency by increasing FTN rates without dire Bit-error-rate sacrifice.

With the aim of addressing power consumption issues for terahertz band wireless communication, this work presents a deep learning-based solution for transceiver design with 1-bit quantization and oversampling at the receiver, and Faster-than-Nyquist transmission over fading channel. Specifically, by implementing the transceiver using a convolutional autoencoder, our work allows higher-order modulation transmission over one-bit fading channel without pilots. Transfer learning from previously trained blocks over simple noisy channel is used to minimize the probability of bit error and outperforms the convolutional autoencoder at low oversampling rates. The biterror-rate gain offered at 20dB SNR by the transfer learning is seen to be as high as half of one order at low oversampling rates and to saturate as oversampling rate increases. Furthermore, by allowing explicit phase synchronization, the autoencoder-based transceiver with partial channel matching is able to approach unquantized performance with 4dB gap in Rayleigh fading environment.

The recent works on a deep learning (DL)-based joint design of preamble set for the transmitters and data-aided active user detection (AUD) in the receiver has demonstrated a significant performance improvement for grant-free sparse code multiple access (GF-SCMA) system. The autoencoder for the joint design can be trained only in a given environment, but in an actual situation where the operating environment is constantly changing, it is difficult to optimize the preamble set for every possible environment. Therefore, a conventional, yet general approach may implement the data-aided AUD while relying on the preamble set that is designed independently rather than the joint design. In this paper, the activity detection error rate (ADER) performance of the data-aided AUD subject to the two preamble designs, i.e., independently designed preamble and jointly designed preamble, were directly compared. Fortunately, it was found that the performance loss in the data-aided AUD induced by the independent preamble design is limited to only 1dB. Furthermore, such performance characteristics of jointly designed preamble set is interpreted through average cross-correlation among the preambles associated with the same codebook (CB) (average intra-CB cross-correlation) and average cross-correlation among preambles associated with the different CBs (average inter-CB cross-correlation).

This paper investigates a joint resource allocation problem in cell-free mobile edge computing system which intends to minimize the number of users subjected to outage, due to failure to meet user-specific delay constraints. Accordingly, the number of APs serving each user, i.e., dynamic cluster size, uplink transmit power and computing resources at the edge server are jointly optimized based on deep reinforcement learning (DRL) algorithm.

In this study, with the aim of reducing power consumption for terahertz band wireless communication, we present a deep learning-based solution for transceiver design with 1-bit quantization and oversampling at the receiver, and Faster- than-Nyquist transmission. Our simulation results illustrate that the studied system with 1-bit quantization achieves bit-error- rate performance comparable to that of an end-to-end channel autoencoder without the constraint of 1-bit quantization subject to the same spectral efficiency. It is also demonstrated that reliable communication can be achieved at rates exceeding 5.3 bits/sec/Hz which corresponds to 80% of the achievable capacity at adequate SNR.

In this paper, with the aim of addressing the stringent computing and quality-of-service (QoS) requirements of recently introduced advanced multimedia services, we consider a cell-free massive MIMO-enabled mobile edge network. In particular, benefited from the reliable cell-free links to offload intensive computation to the edge server, resource-constrained end-users can augment on-board (local) processing with edge computing. To this end, we formulate a joint communication and computing resource allocation (JCCRA) problem to minimize the total energy consumption of the users, while meeting the respective user-specific deadlines. To tackle the problem, we propose a distributed solution approach based on cooperative multi-agent reinforcement learning framework, wherein each user is implemented as a learning agent to make joint resource allocation relying on local information only. The simulation results demonstrate that the performance of the proposed distributed approach outperforms the heuristic baselines, converging to a centralized target benchmark, without resorting to large over-head. Moreover, we showed that the proposed algorithm has performed significantly better in cell-free system as compared with the cellular MEC systems, e.g., a small cell-based MEC system.

Grant-free (GF) access has been one of the enablers for the various use cases in 5th generation (5G) mobile system, especially for time-critical massive machine-type communication (mMTC). However, these use cases have diverse quality of service (QoS) requirements, which can be measured in terms of an access success rate from a GF random access perspective. Consequently, a GF scheme that enables supporting of diverse QoS is highly sought. This article proposes a GF access scheme in which high-QoS users superpose multiple preambles to improve their access success rate as a result of the diversity in access collision and multiple access interference seen by multiple preambles. We further show that in the presence of multiple antennas in the base station (BS), a low-complexity receiver can correctly detect active preambles with a significantly high probability, even under severe multiple access interference caused by non-orthogonal preamble transmission. A theoritical performance analysis is conducted by modeling the preamble reception as a multiple measurement vector-based compressive sensing problem. The preamble misdetection probability is shown to decrease exponentially as the number antennas at BS increases. Numerical results demonstrate multiple-order improvement in terms of the access success rate for critical-QoS users, even under severe noise and multiple access contamination.

The performance of grant-free random access (GF-RA) is limited by the number of accessible random access resources (RRs) due to the absence of collision resolution. Compressive sensing (CS)-based RA schemes scale up the RRs at the expense of increased non-orthogonality among transmitted signals. This paper presents the design of multi-sequence spreading random access (MSRA) which employs multiple spreading sequences to spread the different symbols of a user as opposed to the conventional schemes in which a user employs the same spreading sequence for each symbol. We show that MSRA provides code diversity, enabling the multi-user detection (MUD) to be modeled into a well-conditioned multiple measurement vectors (MMVs) CS problem. The code diversity is quantified by the decrease in the average Babel mutual coherence among the spreading sequences. Moreover, we present a two-stage active user detection (AUD) scheme for both wideband and narrowband implementations. Our theoretical analysis shows that with MSRA activity misdetection falls exponentially while the size of GF-RA frame is increased. Finally, the simulation results show that about 82% increase in utilization of RRs, i.e., more active users, is supported by MSRA than the conventional schemes while achieving the RA failure rate lower bound set by random access collision.

In this paper, the concept of faster-than Nyquist (FTN) transmission is used as a non-orthogonal multiple access scheme to interlace sporadic traffic into a broadband service traffic so that radio resources can be efficiently shared by both traffics, beyond the Nyquist rate. We have extended the concept of a 2-dimensional faster- than Nyquist (2D-FTN) transmission where some time-frequency resources are offloaded (turned off) from a broadband transmission to a time- frequency- space 3DFTN system. A broadband user transmitting in multiple- input multiple-output (MIMO) mode turns-off some set of transmit antennas for few subcarriers, which will be allowed for sporadic traffic. Furthermore, we show that turning-off subcarriers with the ill- conditioned channel matrices reduces the performance penalty of FTN transmission. For example, we have have shown that turning-off 12.5% of subcarriers from broadband transmission of 256 OFDM-symbol block costs only 0.3dB transmit power per bit while allowing up to 64 single- carrier grant-free transmission over the offloaded subcarriers.

In this paper, we introduce a comprehensive grant-free random access scheme for machine-type communication which is characterized by massive connectivity and low latency. The scheme presented in here is comprehensive in a sense that, synchronization, channel estimation, and users identification & data detection (multi-user detection) are performed all in a single shot. The scheme employs compressive sensing by exploiting two sparse phenomena: sparsity in users activity and sparsity in multi-path channel. Furthermore, the scheme is designed in such a way that channel estimation and multi-user detection have a bi-directional mutual relationship, enabling one to reinforce the other for accurate detection and estimation. Moreover, the iterative order recursive least square (IORLS) estimation algorithm is modified & employed in such a way that it exploits the joint structure in multi-path channel and multi-user signal sparsity.

The compressive sensing-based random access may get impacted largely by other cell interference (OCI), which may shatter the sparsity of the received signal model unless the signature space assignment among interfering cells is carefully handled. In this paper, we investigate the performance of multi-sequence spreading-based random access (MSRA) scheme under multi-cell environment. In particular, a grant-free random access scenario is considered to indicate its particular attributes that fundamentally affect the cellular performance. We also give the performance of MSRA in comparison with sparse code multiple access (SCMA). We show that if a proper signature assignment is undertaken, the OCI in MSRA and other CS-based schemes can be modeled as a dispersed noise which does not critically undermine the underlying sparsity.