In this article, a framework of artificial intelligence (AI)-native cross-module optimized physical layer with cooperative control agents is proposed, which involves optimization across global AI/machine learning (ML) modules of the physical layer with innovative design of multiple enhancement mechanisms and control strategies. Specifically, it achieves simultaneous optimization across global modules of uplink AI/ML-based joint source-channel coding with modulation, and downlink AI/ML-based modulation with precoding and corresponding data detection, reducing traditional inter-module information barriers to facilitate end-to-end optimization toward global objectives. Moreover, multiple enhancement mechanisms are also proposed, including i) an AI/ML-based cross-layer modulation approach with theoretical analysis for downlink transmission that breaks the isolation of inter-layer features to expand the solution space for determining improved constellation, ii) a utility-oriented precoder construction method that shifts the role of the AI/ML-based CSI feedback decoder from recovering the original CSI to directly generating precoding matrices aiming to improve end-to-end performance, and iii) incorporating modulation into AI/ML-based CSI feedback to bypass bit-level bottlenecks that introduce quantization errors, non-differentiable gradients, and limitations in constellation solution spaces. Furthermore, AI/ML-based control agents for optimized transmission schemes are proposed that leverage AI/ML to perform model switching according to channel state, thereby enabling integrated control for global throughput optimization. Finally, simulation results demonstrate the superiority of the proposed solutions in terms of block error rate and throughput. These extensive simulations employ more practical assumptions that are aligned with the requirements of the 3rd Generation Partnership Project (3GPP), which hopefully provides valuable insights for future 3GPP standardization discussions.

Accurate and effective channel state information (CSI) feedback is a key technology for massive multiple-input and multiple-output systems. Recently, deep learning (DL) has been introduced for CSI feedback enhancement through massive collected training data and lengthy training time, which is quite costly and impractical for realistic deployment. In this article, a knowledge-driven meta-learning approach is proposed, where the DL model initialized by the meta model obtained from meta training phase is able to achieve rapid convergence when facing a new scenario during target retraining phase. Specifically, instead of training with massive data collected from various scenarios, the meta task environment is constructed based on the intrinsic knowledge of spatial-frequency characteristics of CSI for meta training. Moreover, the target task dataset is also augmented by exploiting the knowledge of statistical characteristics of wireless channel, so that the DL model can achieve higher performance with small actually collected dataset and short training time. In addition, we provide analyses of rationale for the improvement yielded by the knowledge in both phases. Simulation results demonstrate the superiority of the proposed approach from the perspective of feedback performance and convergence speed.

Accurate and effective channel state information (CSI) feedback is a key technology for massive multiple-input and multiple-output (MIMO) systems. Recently, deep learning (DL) has been introduced to enhance CSI feedback in massive MIMO application, where the massive collected training data and lengthy training time are costly and impractical for realistic deployment. In this paper, a knowledge-driven meta-learning solution for CSI feedback is proposed, where the DL model initialized by the meta model obtained from meta training phase is able to achieve rapid convergence when facing a new scenario during the target retraining phase. Specifically, instead of training with massive data collected from various scenarios, the meta task environment is constructed based on the intrinsic knowledge of spatial-frequency characteristics of CSI for meta training. Moreover, the target task dataset is also augmented by exploiting the knowledge of statistical characteristics of channel, so that the DL model initialized by meta training can rapidly fit into a new target scenario with higher performance using only a few actually collected data in the target retraining phase. The method greatly reduces the demand for the number of actual collected data, as well as the cost of training time for realistic deployment. Simulation results demonstrate the superiority of the proposed approach from the perspective of feedback performance and convergence speed.

Beamforming (BF) is usually adopted in millimeter wave wireless systems to compensate high path loss. To increase the received power at user equipment (UE), beam training (BT) is performed to align the beam direction between base station (BS) and UE, where good trade-off between BT overhead and BF gain is preferred. In this letter, two memory-less statistical BT algorithms with finite slots are proposed to reduce the overhead and failure probability of BT, including minimum BT overhead (MO-BT) algorithm and minimum failure probability of BT (MFP-BT) algorithm. Both algorithms make the assumption that memory-less BS randomly select a beam in each slot, based on the prior probability of UE preference, and UE feeds back the beam index to BS to inform which beam should be adopted. By carefully adjusting the probability mass function of beam scanning at BS, MFP-BT can reduce the failure probability of BT, and MO-BT can reduce BT time consumption when compared to their traditional BT counterpart. Numerical results demonstrate the superior performance of our proposed algorithms.

Reconfigurable intelligent surfaces (RISs), which can deliberately adjust the phase of incident waves, have shown enormous potentials to reconFigure the signal propagation for performance enhancement. In this paper, we investigate the RIS-aided offshore system to provide a cost-effective coverage of high-speed data service. The shipborne RIS is placed offshore to improve the signal quality at the vessels, and the coastal base station is equipped with low-cost reconfigurable reflect-arrays (RRAs), instead of the conventional costly fully digital antenna arrays (FDAAs), to reduce the hardware cost. In order to meet the rate requirements of diversified maritime activities, the effective sum rate (ESR) is studied by jointly optimizing the beamforming scheme and the service time allocated to each vessel. The optimal allocation scheme is derived, and an efficient fixed-point based alternating ascent method is developed to obtain a suboptimal solution to the non-convex beamforming problem. Numerical results show that the ESR is considerably improved with the aid of the RIS, and the proposed scheme using the hardware-efficient RRAs has only a slight performance loss, compared to its FDAA-based counterpart.

In millimeter-wave communications, full-dimension multiple-input multiple-output has been widely utilized to generate both horizontal and vertical beams by equipping uniform planar array (UPA) at the base station. However, considering the antenna radiation pattern for each sector, UPA offers nonuniform coverage in horizontal domain, where UEs on the boundary of each sector possess lower beamforming gains and hence the loss of spectral efficiency compared to the UEs at the center. In this paper, uniform cylindrical array (UCA) based hybrid precoding is proposed to improve the uniformity of beam coverage in horizontal domain and the fairness of spectral efficiency for UEs at different locations. Firstly, the UCA-based codebook is designed to generate uniform beams in horizontal domain. Next, the sector-free sub- connection transmitter architecture is designed to reduce the hardware complexity. Meanwhile, to serve multi-user transmission, the corresponding cyclic UE scheduling along with grouped digital precoding are introduced to suppress the inter-user interference. Theoretical analyses and simulation results demonstrate that the proposed scheme can improve the fairness of spectral efficiency compared to its conventional UPA counterpart.

For multi-user millimeter wave (mmWave) communications, different users can be served using various timefrequency resources through their corresponding preferred beams which can provide the highest beamforming gain among all candidate beams. However, when the non-uniform distribution of users is considered, beams selected by fewer users can provide more time-frequency resources to each user than the beams serving more users, which indicates that the conventional uniform beam pattern may lead to unfair resource allocation and hence the loss of system spectral efficiency. In this paper, a novel non-uniform beam design is proposed, wherein a two-level codebook which supports both wide and sharp beams is adopted for data transmission. Specifically, two neighboring sharp beams selected by fewer users can be merged into one wide beam which can cover the same angular space. Thus, the saved resources can be assigned to other sharp beams selected by relatively more users. Theoretical analysis and numerical simulations demonstrate the superiority of our proposal on fairness of resource allocation as well as the spectral efficiency compared with its conventional counterpart.

Due to the inherent high path-loss of millimeterwave signals, beamforming combined with massive MIMO is crucial for increasing coverage range, whereby beam alignment via beam training is necessary to select appropriate beamforming/combining vectors at both base station (BS) and user equipment (UE) in order to acquire sufficient beamforming gain. To reduce the beam training overhead and keep the good beam alignment for millimeter-wave systems, a calibrated beam training is proposed in this paper. Firstly, BS provides a wide beam for coarse beam training and UE calculates the wide beam index by checking the received power levels. Secondly, UE refines the beam training by calculating the narrow beam index based on the ratio of beamforming gain between the selected wide beam and its neighboring wide beams. Simulation results demonstrate that the proposed scheme achieves almost the same beamforming gain as the optimal exhaustive search scheme with much reduced beam training overhead.

Hybrid analog/digital precoding provides an effective solution for millimeter wave (mmWave) multi-input multi-output (MIMO) systems since it reduces the number of radio frequency (RF) chains significantly. Meanwhile, full-dimension MIMO (FD-MIMO) has attracted considerable interests due to its spatial separation ability in vertical direction. However, the hardware complexity of phase shifter (PS) feeding network in the conventional full-connection architecture is too large. In this paper, we propose a novel two-layer PS feeding network architecture for mmWave FD-MIMO systems. Specifically, the two-layer PS network consists of a vertical PS layer and a horizontal PS layer, which greatly reduces the number of PSs required. Furthermore, an iterative precoding design algorithm is proposed, in which the vertical and horizontal beamformers are alternately configured in an iterative fashion. Asymptotic analysis proves that our proposed scheme could exactly approach the performance of full- connection architecture when vertical channel angle spread tends to zero. Simulation results demonstrate that the proposed scheme achieves satisfactory performance while only imposing relatively low hardware complexity compared with its conventional counterparts.

Full-dimensional multiple-input multiple-output (FD-MIMO) systems, whereby each base station is equipped with a uniformly spaced rectangular antenna array (URA), provides a practical means of realizing massive MIMO systems. However, the spectral efficiency of URA is considerably lower than that of its uniformly spaced linear array counterpart having the same number of antenna elements. In this paper, we first introduce a discrete angular resolution metric for quantifying the low resolution of URA in the antenna-elevation domain. This motivates us to propose a novel antenna device design, referred to as the structured non-uniformly spaced rectangular array (NURA), in which the antenna elements are non-uniformly distributed in the elevation-angle domain. Specifically, we conceive a structured NURA device for which the nonuniform distribution of the elevation-domain antenna elements is controlled by a single parameter. The design of the optimally structured NURA for the given nonlinear antenna-element-positioning function then becomes a single-parameter optimization, namely, that of maximizing the spectral efficiency of the FD-MIMO system, which can be solved efficiently. Our simulation results demonstrate that our structured NURA design significantly outperforms the standard URA in terms of achievable spectral efficiency. Our proposed structured NURA design therefore offers an effective practical framework for enhancing the achievable performance of FD-MIMO systems.

In this paper, a two-dimensional location-based channel estimation algorithm is proposed for massive full-dimensional multiple-input multiple-output (FD-MIMO) systems with intracell pilot reuse, which serves more users using the same timefrequency resource and significantly improves the sum capacity of users compared with conventional schemes. By utilizing the two-dimensional antenna arrays, different users assigned with the same pilot could be distinguished by their non-overlapping azimuth angle-of-arrivals (A-AOAs) and elevation angle-of-arrivals (EAOAs), and thus the interference caused by pilot reuse could be removed. Simulation results show that the sum capacity of the proposed algorithm could be improved as the number of antennas grows large.