The channel state information (CSI) acquisition plays a pivotal role in cell-free (CF) massive multiple-input-multi-output (mMIMO) systems. However, conventional pilot-based channel estimation incurs prohibitive overhead costs as user density and mobility increase. To address this, we propose a multi-slot alternating estimation–prediction (MAEP) framework, which leverages temporal correlation to predict future CSI directly and thereby drastically reduce pilot overhead. The efficacy of the proposed framework hinges on prediction accuracy. Inspired by the remarkable modeling capabilities of large language models (LLMs) and their demonstrated efficacy in cross-modal applications, we introduce an LLM-driven channel predictor termed frequency-temporal alignment with LLM (FTAlign-LLM). FTAlign-LLM bridges the modality gap between CSI and the LLM’s feature space through three key components:(i) a multi-scale CSI attention (MSCA) network for extracting rich spatiotemporal features across frequency and delay domains, (ii) a frequency–temporal feature fusion (FTFF) network that fuses these features and aligns them with the LLM’s feature space, and (iii) the utilization of parameter-efficient fine-tuning for LLM adaptation. Extensive results demonstrate that FTAlign-LLM significantly outperforms benchmarks in prediction accuracy. Concurrently, the MAEP framework achieves substantial improvements in sum spectral efficiency, particularly when a large number of access points are deployed in CF mMIMO systems.

Semantic communication (SemCom) has recently emerged as a transformative paradigm for next-generation wireless systems, aiming to enable task-oriented information transmission. However, existing SemCom networks rely on cell-centric architectures, where users, particularly those at the cell edge, suffer from severe inter-cell interference and poor channel conditions. These issues undermine semantic accuracy, induce intolerable delay, and ultimately hinder the fulfillment of quality-of-service (QoS) requirements. To overcome these limitations, we propose the user-centric SemCom (UCSC) network, a novel architecture that integrates user-centric networks with semantic-aware transmission. In UCSC, each user is served by a dedicated cooperative access point group, enabling effective interference mitigation and reliable semantic delivery, even at the cell edge. To fully exploit the potential of UCSC, we formulate a joint optimization problem for semantic information extraction and resource allocation, aiming to minimize total energy consumption under constraints on semantic similarity and delay. By decomposing the original problem into three tractable subproblems, semantic information extraction, computational capacity control, and power allocation, we propose an alternating optimization algorithm to efficiently solve it. Simulation results demonstrate that UCSC, combined with the proposed algorithm, achieves significant performance improvements in reducing energy consumption compared with cell-centric SemCom networks, especially under stringent QoS requirements.

The limited-capacity fronthaul link between access points (APs) and central processing unit (CPU), along with signal distortions caused by hardware impairments at users and APs, lead to unreliable transmission, hindering the achievement of ultra-reliable low-latency communication (URLLC) in cell-free (CF) massive multiple-input-multiple-output (mMIMO) systems. In this paper, we investigate the performance of URLLC in CF mMIMO systems with limited fronthaul capacity and hardware impairments. First, a low-complexity fronthaul rate allocation scheme is proposed using rate-distortion theory in the considered system. Then, we derive the distribution of signal-to-interference-plus noise ratio (SINR) and analyze the delay violation probability (DVP) based on stochastic network calculus. Furthermore, a closed-form upper bound expression for DVP is also derived. To fulfill the requirements of next-generation URLLC, which demand additional key performance indicators, we further investigate effective energy efficiency and derive a closed-form lower bound, defined as the ratio of effective capacity to total power consumption. Extensive numerical results verify the accuracy of our derivations and indicate that a large deployment of APs can compensate for the reliability degradation caused by limited fronthaul capacity and hardware impairments. Additionally, the number of APs must be appropriately configured to ensure low latency and high reliability while avoiding energy waste.

Cell-free (CF) massive multiple-input multiple-output (mMIMO) has been considered as a potential technology for Beyond 5G. However, the performance of CF mMIMO systems has not been thoroughly studied. Most existing analytical studies on CF mMIMO systems rely on deriving average performance metrics. The statistical characteristics of the signal-to-interference-plus-noise ratio (SINR), which capture the tail behavior of SINR, are crucial for metrics such as outage probability and for emerging mission-critical applications that emphasize extreme and rare events, but have not been thoroughly investigated. In this paper, we aim to obtain the distribution of SINR in CF mMIMO systems. Considering a downlink CF mMIMO system with pilot contamination, we first give the closed-form expression of the SINR. Based on our analytical work on the two components of the SINR, i.e., desired signal and interference-plus-noise, we then derive the probability density function and cumulative distribution function of the SINR under maximum ratio transmission (MRT) and full-pilot zero-forcing (FZF) precoding, respectively. Subsequently, the closed-form expressions for two more sophisticated performance metrics, i.e., ergodic rate and outage probability, are obtained. Finally, we perform Monte Carlo simulations to validate our analytical work. Numerous numerical results demonstrate the effectiveness of the derived SINR distribution, ergodic rate, and outage probability.

Ultra-reliable low-latency communication (URLLC) has emerged as a key service class for enabling next-generation mission-critical applications. However, the increasingly stringent quality-of-service requirements pose significant challenges for URLLC under traditional bit-level communication (BitCom), especially in handling rare and extreme events that are likely to occur in low signal-to-noise ratio (SNR) conditions and resource-constrained wireless environments. Semantic communication (SemCom), by compressing source information toward task-relevant content, demonstrates clear advantages over BitCom in low SNR regimes and under limited resource availability. Furthermore, SemCom inherently conveys task-specific meanings, making it well-suited for mission-critical applications in machine-to-machine communications. Motivated by these observations, we investigate the performance of URLLC under the SemCom paradigm, and derive a closed-form upper bound on the delay violation probability (DVP) using stochastic network calculus. To further exploit the strengths of both SemCom and BitCom, we propose a hybrid semantic and bit communication paradigm (HSBP). To ensure URLLC performance under HSBP, we first conduct a theoretical performance analysis and, based on this analysis, propose a bisection-based bandwidth allocation algorithm to minimize the maximum DVP. Extensive simulation results validate the accuracy of the derived upper bound and demonstrate that the proposed HSBP paradigm outperforms pure SemCom and pure BitCom in terms of DVP.

Cell-free (CF) massive multiple-input multiple-output (mMIMO), which has been considered as a promising technology to achieve ultra-reliable low latency communications (URLLC), emphasizes extreme and rare events instead of well studied average performance. In this paper, we analyze the statistical delay performance in uplink CF mMIMO-aided URLLC systems, specifically characterizing the tail of the delay distribution. Firstly, we derive the SINR distribution in a user-centric uplink CF mMIMO system using moment-matching techniques and log-normal approximation. Subsequently, the average decoding error probability (DEP) with finite blocklength coding is studied based on the derived distribution, and an upper bound on the delay violation probability (DVP) is analyzed using stochastic network calculus (SNC). To provide a clearer representation of the bound on DVP, a closed-form upper bound for the average DEP is derived. Furthermore, we propose a rate adaptive scheme based on SNC to minimize the DVP. Numerous numerical results validate the accuracy of the derived SINR distribution and the upper bound on average DEP. Besides, CF mMIMO systems exhibit significant gains compared to mMIMO systems in terms of statistical delay performance, as the distributed deployment of numerous APs can effectively compensate for performance degradation caused by interference and increased arrival rates.

User-centric strategies in cell-free (CF) massive multiple-input multiple-output (mMIMO)-aided ultra-reliable low-latency communication (URLLC) systems can significantly reduce backhaul pressure to ensure low latency. Furthermore, appropriate AP selection algorithms can guarantee reliable QoS support for URLLC. However, existing AP selection algorithms in CF mMIMO-aided URLLC systems are designed solely based on large-scale fading information, which is a heuristic algorithm that makes it difficult to guarantee performance. In this paper, we consider an uplink user-centric CF mMIMO-enabled URLLC system and formulate a weighted sum rate maximum problem by jointly optimizing AP selection and power control. Since the problem is non-convex, we first decompose the original problem into a series of subproblems by the log-function transformation and fractional programming. Subsequently, we divide each subproblem into two parts for iterative solutions: one part deals with AP selection with fixed power allocation which can be solved using linear conic relaxation, and the other part involves power control with fixed AP allocation, and can be directly solved. Finally, the simulation results validate that our proposed algorithm converges rapidly, and has significant performance improvement compared to the benchmark algorithms.

Serverless computing is gaining popularity due to its "pay-per-use" model and simplified management. Benefiting from the common features of Internet of Things (IoT), the adaptation of serverless in edge computing is extensively researched. To protect nodes in serverless edge computing from potential attacks, blockchain technology plays a crucial role. Considering that the computing time of a single serverless function can be less than the time for a blockchain to generate a block, the blockchain component requires significant computing and bandwidth resources to ensure the generation efficiency of blocks matches the execution efficiency of serverless functions. However, existing studies on minimizing the completion time (a.k.a. makespan) of application workflows in serverless edge computing have never considered the impact brought by blockchain, leading to an imbalance in the resource allocation between the computing and blockchain components. To address this, we propose a Heterogeneous Event-driven Scheduling (HEDS) mechanism, which employs a fine-grained CPU al-location scheme and an event-driven approach, to make the adjustment of resource allocation between the computing and blockchain components more flexible and expeditious, leading to improving the utilization of both computing and bandwidth resources. Meanwhile, a Multi-agent Double Deep Q-network (MADDQN) algorithm is proposed to help agents minimize the makespan while balancing the resource consumption in the computing and blockchain components. Simulation results demonstrate that HEDS with MADDQN outperforms existing algorithms in terms of average makespan over different numbers of functions and nodes while satisfying the latency requirements of block generation.

Integrated sensing and communication (ISAC) is an emerging technology in next-generation communication networks. However, the communication performance of the ISAC system may be severely affected by interference from the radar system if the sensing task has demanding performance requirements. In this paper, we exploit device-to-device (D2D) communication to improve communication capacity of the ISAC system. The ISAC system in a single cell D2D assisted-network is investigated, where the base station (BS) performs target sensing and communication with multiple cellular user equipments (CUEs) as well as D2D user equipments (DUEs) simultaneously communicating with other DUEs by multiplexing the same frequency resource. To achieve the optimal communication performance in such D2D-assisted ISAC system, a joint beamforming and power control problem is formulated with the goal to maximize the sum rate while guaranteeing the performance requirements of radar sensing. Due to the non-convexity of the problem, we transform the origin problem into a relaxation form, then, a joint beamforming and D2D power control algorithm is proposed to obtain the solution efficiently. Particularly, a zero-forcing (ZF) beamforming scheme to eliminate the interference from the BS on DUEs is also proposed. Extensive numerical simulations demonstrated that with the assistance of the D2D communications, our proposed algorithm significantly outperforms the baseline schemes in terms of the system sum rate.

Cell-free (CF) massive multiple-input multiple-output (mMIMO), characterized by macro-diversity and spatial sparsity, has been considered as a potential technology to support ultra-reliable low-latency communication (URLLC). The average performance has been comprehensively investigated for URLLC in CF mMIMO systems. However, URLLC places its central focus on extreme and rare events, requiring statistical quality of service (QoS) provisioning in CF mMIMO systems. In this paper, we model the statistical QoS provisioning constraints for URLLC in a CF mMIMO system based on extreme value theory (EVT), i.e., delay violation probability boundary and statistical properties of extreme queue values. Based on our analytical work, a power control optimization problem with long-term URLLC constraints is formulated, aiming at minimizing energy consumption. Then, Lyapunov optimization is utilized to decompose this long-term stochastic optimization problem into a series of short-term deterministic problems. Since the short-term problems are non-convex and intractable, a learning-based hyper-heuristic algorithm, consisting of a high-level strategy and multiple low-level heuristics, is proposed. Numerical results verify the effectiveness of parameterizing URLLC in the CF mMIMO system based on EVT and demonstrate that the proposed algorithm outperforms benchmark algorithms in both average delay and delay fluctuations, achieving statistical QoS provisioning.

In the traditional cellular-based mobile edge computing (MEC), users at the edge of the cell are prone to suffer severe inter-cell interference and signal attenuation, leading to low throughput even transmission interruptions. Such edge effect severely obstructs offloading of tasks to MEC servers. To address this issue, we propose user-centric mobile edge computing (UCMEC), a novel MEC architecture integrating user-centric transmission, which can ensure high throughput and reliable communication for task offloading. Then, we formulate an long-term delay minimization problem by jointly optimizing task offloading, power allocation, and computing resource allocation in UCMEC. To solve the intractable problem, we propose two decentralized joint optimization schemes based on multi-agent deep reinforcement learning (MADRL) and convex optimization, which consider both cooperation and non-cooperation among network nodes. Simulation results demonstrate that the proposed schemes in UCMEC can significantly improve the uplink transmission rate by at least 176.99% and reduce the long-term average total delay by at least 16.36% compared to traditional cellular-based MEC.

Inter-user interference and pilot contamination are the major obstacles limiting the performance of cell-free massive multiple-input multiple-output (CF mMIMO)-aided ultra-reliable low-latency communication (URLLC) systems. In this paper, user grouping is utilized to address these issues, by allocating users to groups based on frequency band division, preventing interference among users within different groups, and eliminating pilot contamination between different groups. We consider an uplink CF mMIMO-aided URLLC system with user grouping and derive the lower bound for the ergodic rate. Due to the limited blocklength of each group necessitating pilot reuse, a weight sum rate (WSR) maximum problem is formulated by jointly optimizing user grouping, pilot assignment, and power control. We propose a user grouping scheme based on graph theory, where iteratively searching for specific negative loops in the weighted directed graph can approach the optimal user grouping matrix. As the user grouping matrix updates at each iteration, we update the pilot assignment matrix based on graph theory and employ logarithmic function approximation and fractional programming for power control updates. Numerous numerical results demonstrate the effectiveness of user grouping, and the proposed algorithm improves WSR by 25% compared to the non-grouping algorithm, while outperforming other benchmark algorithms.

Segment Routing over IPv6 (SRv6) is a promising efficient technology for traffic engineering (TE). As transitioning from a traditional distributed network to a full SRv6 network faces technical and economic challenges, partially deploying SRv6 has attracted much attention from academic communities. Many TE research attempts have been made on SRv6 deployment and routing optimization, among which Deep Reinforcement Learning (DRL) based algorithms have shown their advantages over traditional algorithms. However, with the incremental deployment of SRv6 nodes, training costs for DRL as well as solution space of routing optimization increase significantly, which obstructs the application of DRL-based algorithms in practice. To address this issue, we propose a DRL-based SRv6 deployment and routing optimization (SDRO) algorithm, with the TE objective of minimizing the maximum link utilization. In SDRO, the DRL agent is only trained once on a full SRv6 network and then used for different SRv6 deployment ratios. Hence, training costs can be obviously reduced. To reduce the solution space of routing optimization, the DRL agent performs routing pre-optimization on a portion of the traffic before routing is finally optimized by the Linear Programming method. By doing so, the execution time for routing optimization can be greatly reduced. Besides, for the issue of frequent link failures in the network, SDRO leverages the generalization of Graph Neural Networks to improve its robustness. Simulation results demonstrate that SDRO outperforms existing algorithms under different SRv6 deployment ratios and link failures, and completes routing optimization in a few seconds.