Integrated sensing and communication (ISAC) is a key enabling technology for sixth-generation (6G) mobile communication systems, achieving seamless integration of communication and sensing functions. Cooperative sensing, where the transmitter and receiver are not co-located, serves as a key enabler for ISAC, significantly enhancing the sensing performance while reducing the implementation complexity of the receiver. However, practical deployments of cooperative sensing still face numerous challenges, such as synchronization and interference. This paper presents a set of advanced beam management methods specifically designed for the cooperative sensing system, offering a comprehensive framework to address these challenges. Specifically, we first analyze the strong/weak path effect (SWPE), a critical phenomenon caused by diverse target reflectivities and propagation paths, which severely degrades both synchronization accuracy and target detection. To counteract this, we propose an adaptive path power allocation method compatible with both all-digital and hybrid beamforming architectures. This method intelligently allocates power across different paths to mitigate the SWPE, thereby ensuring reliable synchronization via the direct path while enhancing the detectability of weak targets. As a result, the proposed method improves the target detection probability by over 30%. Furthermore, an adaptive interference suppression method is designed to reduce interference while maintaining sensing/communication quality, which obtains the SINR gain of around 5 dB, compared to the traditional full nulling method. Experimental results validate the effectiveness of robust synchronization and our proposed power allocation. This study lays a solid foundation for beamforming optimization in cooperative sensing systems, facilitating high-accuracy sensing and communication in complex environments.

Channel estimation in a fading environment can be regarded as a typical statistical estimation problem. Its optimal performance relies on the prior distribution of the channel coefficients, which are environment-specific. However, the conventional channel estimators, such as the least square (LS) and linear minimum mean square error (LMMSE) estimators, do not fully exploit the prior channel distribution law. To address this limitation, we propose to use the conditional diffusion model (DM) to achieve environment-aware channel estimation, referred to as EACE-DM. In this framework, the environment information is incorporated as the condition to guide the EACE-DM in learning the hidden features of channels from various environments. The trained EACE-DM is functionally decomposed into two components, i.e., the environment identification module and the channel estimation module. The environment identification module first uses the DM’s forward process to diffuse LS channel estimation into a noisy sample. Then it executes the DM’s reverse denoising process conditioned on candidate environments to recover the LS estimation from the noisy channel. The environment is then identified via maximum a posteriori (MAP) estimation by comparing these recovered estimations with the ground-truth LS estimation. Finally, the identified environment is utilized to guide the channel estimation module, denoising the LS estimation. Numerical simulations demonstrate that the proposed EACE-DM significantly decreases normalized mean square errors (NMSEs) of channel estimation across diverse environments while incurring a moderate increase in computational complexity compared to conventional estimators and existing DM-based approaches.

Deep learning is promising to enhance the accuracy and reduce the overhead of channel state information (CSI) feedback, which can boost the capacity of frequency division duplex (FDD) massive multiple-input multiple-output (MIMO) systems. Nevertheless, the generalizability of current deep learning-based CSI feedback algorithms cannot be guaranteed in unseen environments, which induces a high deployment cost. In this paper, the generalizability of deep learning-based CSI feedback is promoted with physics interpretation. Firstly, the distribution shift of the cluster-based channel is modeled, which comprises the multi-cluster structure and single-cluster response. Secondly, the physics-based distribution alignment is proposed to effectively address the distribution shift of the cluster-based channel, which comprises multi-cluster decoupling and fine-grained alignment. Thirdly, the efficiency and robustness of physics-based distribution alignment are enhanced. Explicitly, an efficient multi-cluster decoupling algorithm is proposed based on the Eckart–Young-Mirsky (EYM) theorem to support real-time CSI feedback. Meanwhile, a hybrid criterion to estimate the number of decoupled clusters is designed, which enhances the robust-ness against channel estimation error. Fourthly, environment-generalizable neural network for CSI feedback (EG-CsiNet) is proposed as a novel learning framework with physics-based distribution alignment. Based on extensive simulations and sim-to-real experiments in various conditions, the proposed EG-CsiNet can robustly reduce the generalization error by more than 3 dB compared to the state-of-the-arts.

In scenarios where the number of users far exceeds the number of base station antennas, it becomes infeasible to serve all users simultaneously. We observe that scheduling more users at first steadily improves system performance, but this improvement eventually reaches saturation. Moreover, scheduling excessive users per time slot results in diminishing gains per user and a significant increase in computational complexity. Consequently, only a subset of users can be selected for service in each time slot. Given that the fairness scheduling and precoding problems are coupled, we employ learning-based methods to rapidly schedule users, reducing the complexity of subsequent joint optimization problems. To facilitate effective backpropagation despite discrete scheduling actions, we design the network's output layer accordingly and introduce a layer-wise pruning mechanism to enhance learning accuracy and speed. By leveraging the interference relationships, we partition the base station into clusters, thereby transforming the centralized joint optimization problem into multiple sub-problems, which can be tackled in a distributed manner.

Efficient power allocation and interference management are critical challenges in dynamic wireless communication systems. To address these challenges, graph neural networks (GNNs) have attracted significant attention, while knowledge graph further enhance this capability by representing structured interactions among entities. This article proposes the Power-focused Knowledge Graph Convolutional Network (PKGCN), a novel framework utilizing knowledge graph driven learning to model and optimize power allocation strategies. By integrating wireless-specific features such as channel conditions and interference metrics, PKGCN effectively captures the complex interactions and dependencies among network nodes. This model employs a message aggregation layer to extract local and global interactions and a power prediction layer to optimize resource allocation. Comprehensive evaluations reveal that PKGCN de-livers higher average user rates, lower interference levels, and greater robustness.

The massive scale of deployment, hundreds of parameters, differentiated scenarios and interworking with existing mobile networks leads to high complexity and high cost of optimization, operation and maintenance of the 5th generation mobile network (5G), which inspires that 6th generation mobile network (6G) should support high level autonomy at the beginning of deployment. Digital twin network (DTN) technology, with its advantages of intelligent decision making, low-cost experimentation, and preverification, has emerged as a key enabling technology for autonomous network. To address the need for flexibility to fulfill more diverse scenarios and high-level autonomy toward 2030, this article discusses the typical usage cases of DTN, and proposes an innovative and native design for 6G DTN, encompassing logical framework, architecture, functions, and deployment modes. Furthermore, the efficient DTN Model Construction and Intelligent Orchestration and Management are introduced to enable fully automated and high-performance DTN tasks. Finally, the future direction for DTN research is presented.

Extended reality (XR), blending virtual and real worlds, is a key application of future networks. While AI advancements enhance XR capabilities, they also impose significant computational and energy challenges on lightweight XR devices. In this article, we developed a distributed queue model for multitask deep neural network inference, addressing issues of resource competition and queue coupling. In response to the challenges posed by the high energy consumption and limited resources of XR devices, we designed a dual time-scale joint optimization strategy for model partitioning and resource allocation, formulated as a bi-level optimization problem. This strategy aims to minimize the total energy consumption of XR devices while ensuring queue stability and adhering to computational and communication resource constraints. To tackle this problem, we devised a Lyapunov-guided proximal policy optimization algorithm, named LyaPPO. Through numerical results, we show that our LyaPPO algorithm outperforms the baseline algorithms. Specifically, under different maximum local computational capacities, the proposed algorithm decreases 24.29%–56.62% energy compared to the suboptimal baselines.

Designing short Luby transformation (LT) codes with low overhead and good error performance is crucial and challenging for the deployment of vehicle-to-everything networks, which require high reliability, high spectral efficiency, and low latency. In this paper, we investigate the design of globally optimal transmission strategies that consider interactions between feedback for short LT codes using reinforcement learning (RL), where traditional asymptotic analysis based on random graph theory is known to be inaccurate in this context. First, in order to reduce the decoding overhead of short LT codes, we derive the gradient expression for optimizing the degree distribution of LT codes, and propose a RL-based distribution optimization (RL-DO) algorithm for designing short LT codes. Then, to improve the reliability and overhead of LT codes under limited feedback, we model the feedback optimization problem as a Markov decision process, and propose the RL-based joint feedback and distribution optimization (RL-JFDO) algorithm, which aims to design globally-optimal feedback schemes. Simulations show that our methods have lower decoding overhead, error rate, and decoding complexity compared to existing feedback fountain codes.

网络资源的覆盖范围日益广泛, 算力资源也逐渐成为能够提供泛在计算服务的基础设施. 然而, 在广域网络, 底层网络和计算资源缺乏密切的研究或协同设计, 仍然存在计算服务调度缓慢、 数据分发不灵活、 数据传输效率低等问题. 本文提出算力感知网络(CAN)的系统架构设计, 其核心贡献在于引入感知平面来收集、 管理并综合计算和网络的信息. 这样, 感知平面、控制平面和数据平面组成一个闭环控制系统, 增强了整个系统的感知能力、 决策能力和数据转发功能. 为了使能CAN系统, 本文提出三项关键技术: 算力路由、 弹性广播和广域高吞吐传输. 本文以人工智能(AI)模型训练、 推理和离线参数传输为例, 展示CAN的适用性, 并指出未来的一些研究方向.

For frequency division duplex (FDD) systems, downlink channel state information (CSI) feedback is essential. Deep learning-based auto-encoder (AE) structures have shown promise in reducing feedback overhead. However, designing a super-large AE network to handle the CSI of all scenarios is not practical. A more practical approach is to divide the CSI dataset by region/scenario and use multiple simple AE networks. However, this method requires high memory capacity, making it unsuitable for low-end user equipment (UE). In this paper, we propose a new UE-friendly framework based on multi-tasking mode. Our framework, called single-encoder-to-multiple-decoders (S-to-M), uses multi-task-learning to design multiple independent AEs into a joint architecture with a shared encoder that corresponds to multiple task-specific decoders. We also integrate GateNet as a classifier to enable the base station to autonomously select the right task-specific decoder for the subregion. Our experiments on a simulated multi-scenario CSI dataset show that our proposed S-to-M framework outperforms other benchmark modes by significantly reducing model complexity and UE memory consumption.

Channel estimation is challenging for the reconfigurable intelligent surface (RIS)-aided wireless communications. Since the number of coefficients of the cascaded channel among the base station (BS), the RIS, and the user equipment (UE), is the product of the number of BS antennas, the number of RIS elements, and the number of UEs, the pilot overhead can be prohibitively high. In this paper, we propose a two-timescale channel estimation framework to exploit the property that the BS-RIS channel is high-dimensional but quasi-static, while the RIS-UE channel is mobile but low-dimensional. Specifically, to estimate the quasi-static BS-RIS channel, we propose a dual-link pilot transmission scheme, where the BS transmits downlink pilots and receives uplink pilots reflected by the RIS. Then, we propose a coordinate descent-based algorithm to recover the BS-RIS channel. Since the quasi-static BS-RIS channel is estimated less frequently than the mobile channel is, the average pilot overhead can be reduced from a long-term perspective. Although the mobile RIS-UE channel has to be frequently estimated in a small timescale, the associated pilot overhead is low thanks to its low dimension. Simulation results show that the proposed two-timescale channel estimation framework can achieve accurate channel estimation with low pilot overhead.

Three-dimensional (3D) multiple input and multiple output (3D-MIMO) with massive antennas is a key technology to achieve high spectral efficiency and user experienced data rate for the fifth generation (5G) mobile communication system. To implement 3D-MIMO in 5G system, practical constraints on the product design should be considered. This paper proposes a systematic design for the 3D-MIMO product by considering the restrictions of both base band and the hardware, including cost, size, weight, and heat dissipation. The design has been implemented for 2.6-GHz time-division duplex band, and field trials have been conducted for performance validation with practical intercell interference in commercial network. The trial results show that this 3D-MIMO design can meet the spectral efficiency requirement of the 5G enhanced mobile broadband services. The performance gain of 3D-MIMO varies with the traffic load. When the traffic load is heavy, 3D-MIMO can enhance the cell throughput by 4~6.7 times. When the traffic load is low, the performance gain of this 3D-MIMO design decreases. The results from field trial also show that the performance of 3D-MIMO degrades in mobility scenarios, where further enhancement on acquiring instant channel status information are necessary to improve the robustness of 3D-MIMO to mobility.