This work proposes novel approaches that jointly design user equipment (UE) association and power control (PC) in a downlink user-centric cell-free massive multiple-input multiple-output (CFmMIMO) network, where each UE is only served by a set of access points (APs) for reducing the fronthaul signalling and computational complexity. In order to maximize the sum spectral efficiency (SE) of the UEs, we formulate a mixed-integer nonconvex optimization problem under constraints on the per-AP transmit power, quality-of-service rate requirements, maximum fronthaul signalling load, and maximum number of UEs served by each AP. In order to efficiently solve the formulated problem, we propose two different schemes according to the different sizes of the CFmMIMO systems. For small-scale CFmMIMO systems, we present a successive convex approximation (SCA) method to obtain a stationary solution and also develop a learning-based method (JointCFNet) to reduce the computational complexity. For large-scale CFmMIMO systems, we propose a low-complexity suboptimal algorithm using accelerated projected gradient (APG) techniques. Numerical results show that our JointCFNet can yield similar performance and significantly decrease the run time compared with the SCA algorithm in small-scale systems. The presented APG approach is confirmed to run much faster than the SCA algorithm in large-scale systems while obtaining an SE performance close to that of the SCA approach. Moreover, the median sum SE of the APG method is up to about 2.8 fold higher than that of the heuristic baseline scheme.

Non-independent and identically distributed (non-IID) data distribution among clients is considered as the key factor that degrades the performance of federated learning (FL). Several approaches to handle non-IID data, such as personalized FL and federated multitask learning (FMTL), are of great interest to research communities. In this work, first, we formulate the FMTL problem using Laplacian regularization to explicitly leverage the relationships among the models of clients for multitask learning. Then, we introduce a new view of the FMTL problem, which, for the first time, shows that the formulated FMTL problem can be used for conventional FL and personalized FL. We also propose two algorithms FedU and decentrali- zed FedU ( dFedU ) to solve the formulated FMTL problem in communication-centralized and decentralized schemes, respectively. Theoretically, we prove that the convergence rates of both algorithms achieve linear speedup for strongly convex and sublinear speedup of order 1/2 for nonconvex objectives. Experimentally, we show that our algorithms outperform the conventional algorithm FedAvg, FedProx, SCAFFOLD, and AFL in FL settings, MOCHA in FMTL settings, as well as pFedMe and Per-FedAvg in personalized FL settings.

Due to its communication efficiency and privacy-preserving capability, federated learning (FL) has emerged as a promising framework for machine learning in 5G-and-beyond wireless networks. Of great interest is the design and optimization of new wireless network structures that support the stable and fast operation of FL. Cell-free massive multiple-input–multiple-output (CFmMIMO) turns out to be a suitable candidate, which allows each communication round in the iterative FL process to be stably executed within a large-scale coherence time. Aiming to reduce the total execution time of the FL process in CFmMIMO, this article proposes choosing only a subset of available users to participate in FL. An optimal selection of users with favorable link conditions would minimize the execution time of each communication round while limiting the total number of communication rounds required. Toward this end, we formulate a joint optimization problem of user selection, transmit power, and processing frequency, subject to a predefined minimum number of participating users to guarantee the quality of learning. We then develop a new algorithm that is proven to converge to the neighborhood of the stationary points of the formulated problem. Numerical results confirm that our proposed approach significantly reduces the FL total execution time over baseline schemes. The time reduction is more pronounced when the density of access point deployments is moderately low.

Infrastructure-to-vehicle (I2V) and vehicle-to-vehicle (V2V) communications are often combined to extend the connectivity and coverage in the Intelligent Transportation System (ITS) and its applications, e.g., augmented reality, real-time parking management and online shopping. Through multi-hop I2V and V2V communications, requesting vehicles are always connected to road side units (RSUs) even when they do not reside within the RSUs’ coverage range. However, there may be not adequate network resource for several I2V and V2V links when multiple vehicles request services simultaneously. In this paper, we propose a joint frequency scheduling and power control scheme to enhance connectivity in multi-hop I2V/V2V networks. We associate I2V and V2V links with tuple-links, then formulate an NP-hard problem in which a frequency scheduler and a power controller are jointly designed for the tuple-links. The NP-hard problem is decomposed into two separate subproblems by employing the delayed column generation technique. Then, we employ a method for linear programming and a greedy algorithm to address these subproblems. Through numerical experiments with practical parameter settings, we demonstrate the proposed scheme outperforms several existing ones in terms of connectivity enhancement, measured by the service resumption number and average achieved throughput. Furthermore, the efficiency of our scheme is further enhanced when the number of available channels is high, and buffer size equipped to the requesting vehicles is large.

Automotive infotainment systems are expected to be first deployed on highways to service drivers travelling long distances, who are more likely to utilize the infotainment applications. In order to meet the stringent requirements of the infotainment systems, road side units (RSUs) are installed along the highway to facilitate a continuous vehicle-to-infrastructure (V2I) connectivity. Due to the long travelling distance and small coverage of the individual RSU, a more cost-effective solution would be to combine V2I with the vehicle-to-vehicle (V2V) communications to maintain the continuous connectivity. In this paper, we propose a new dynamic cooperation scheme that employs a dynamic forwarder selection strategy to generate an adaptive multi-hop V2V path for connectivity maintenance and throughput enhancement at a vehicle located outside of the RSU’s coverage range. For the commonly assumed scenario that all vehicles travel in the same direction and at the same speed, we develop an analytical model and derive closed-formed expressions for the average out-of-range connection time, number of service resumptions and achieved throughput. The developed analytical model provides insights into the impacts of inter-RSU distance, vehicles’ assistance willingness and the target vehicle’s buffer size to the network performance. Simulation results with practical parameter settings show that our proposed scheme is effective in improving connectivity while offering a high throughput for the target vehicle. In particular, a high vehicle density, more assistance willingness by the forwarders and a large buffer size at the target vehicle are shown to be helpful in sparse RSU deployments.

With its privacy preservation and communication efficiency, federated learning (FL) has emerged as a learning framework that suits beyond 5G and towards 6G systems. This work looks into a future scenario in which there are multiple groups with different learning purposes and participating in different FL processes. We give energy-efficient solutions to demonstrate that this scenario can be realistic. First, to ensure a stable operation of multiple FL processes over wireless channels, we propose to use a massive multiple-input multiple-output network to support the local and global FL training updates, and let the iterations of these FL processes be executed within the same large-scale coherence time. Then, we develop asynchronous and synchronous transmission protocols where these iterations are asynchronously and synchronously executed, respectively, using the downlink unicasting and conventional uplink transmission schemes. Zero-forcing processing is utilized for both uplink and downlink transmissions. Finally, we propose an algorithm that optimally allocates power and computation resources to save energy at both base station and user sides, while guaranteeing a given maximum execution time threshold of each FL iteration. Compared to the baseline schemes, the proposed algorithm significantly reduces the energy consumption, especially when the number of base station antennas is large.

Straggler effect is the main bottleneck in realizing federated learning (FL) in wireless networks. This work proposes a novel user (UE) selection approach to mitigate this effect with UE sampling in cell-free massive multiple-input multiple-output networks. Our proposed approach selects only a small subset of UEs for participating in one FL process. Importantly, since the UEs are selected before any FL process is executed, the performance of FL during the executing time is not affected by our method. Here, we select UEs by solving an FL transmission time minimization problem that jointly optimizes UE selection, power control, and data rate. The problem is formulated to capture the complex interactions among the FL training time, UE selection, and straggler effect. This mixed-integer mixed-timescale stochastic nonconvex problem is constrained by the minimum number of UEs to guarantee the quality of learning. By employing online successive convex approximation, we propose a novel algorithm to solve the formulated problem with guaranteed convergence to the neighbourhood of their stationary points. Our approach can significantly reduce the FL transmission time over baseline approaches, especially in the networks that experience serious straggler effect due to the moderately low density of access points.

This paper proposes a novel scheme for cell-free massive multiple-input multiple-output (CFmMIMO) networks to support any federated learning (FL) framework. This scheme allows each instead of all the iterations of the FL framework to happen in a large-scale coherence time to guarantee a stable operation of an FL process. To show how to optimize the FL performance using this proposed scheme, we consider an existing FL framework as an example and target FL training time minimization for this framework. An optimization problem is then formulated to jointly optimize the local accuracy, transmit power, data rate, and users' processing frequency. This mixed-timescale stochastic nonconvex problem captures the complex interactions among the training time, and transmission and computation of training updates of one FL process. By employing the online successive convex approximation approach, we develop a new algorithm to solve the formulated problem with proven convergence to the neighbourhood of its stationary points. Our numerical results confirm that the presented joint design reduces the training time by up to 55% over baseline approaches. They also show that CFmMIMO here requires the lowest training time for FL processes compared with cell-free time-division multiple access massive MIMO and collocated massive MIMO.

This work aims to maximize the energy efficiency of a downlink cloud radio access network (C-RAN), where data is transferred from a baseband unit in the core network to several remote radio heads via a set of edge routers over capacity-limited fronthaul links. The remote radio heads then send the received signals to their users via radio access links. We formulate a new mixed-integer nonlinear problem in which the ratio of network throughput and total power consumption is maximized. This challenging problem formulation includes practical constraints on routing, predefined minimum data rates, fronthaul capacity and maximum RRH transmit power. By employing the successive convex quadratic programming framework, an iterative algorithm is proposed with guaranteed convergence to a Fritz John solution of the formulated problem. Significantly, each iteration of the proposed algorithm solves only one simple convex program. Numerical examples with practical parameters confirm that the proposed joint optimization design markedly improves the C-RAN's energy efficiency compared to benchmark schemes.

This paper aims to maximize the spectral and energy efficiencies of a content-centric cloud radio access network (C-RAN), where users requesting the same contents are grouped together. Data are transferred from a central baseband unit to multiple remote radio heads (RRHs) equipped with local caches. The RRHs then send the received data to each group's user. Both multicast and unicast schemes are considered for data transmission. We formulate mixed-integer nonlinear problems in which user association, RRH activation, data rate allocation, and signal precoding are jointly designed. These challenging problems are subject to minimum data rate requirements, limited fronthaul capacity, and maximum RRH transmit power. Employing successive convex quadratic programming, we propose iterative algorithms with guaranteed convergence to Fritz John solutions. Numerical results confirm that the proposed joint designs markedly improve the spectral and energy efficiencies of the considered content-centric C-RAN compared to benchmark schemes. Importantly, they show that unicasting outperforms multicasting in terms of spectral efficiency in both cache and cache-less scenarios. In terms of energy efficiency, multicasting is the best choice for the system without cache whereas unicasting is best for the system with cache. Finally, edge caching is shown to improve both spectral and energy efficiencies.

This paper aims to maximize the energy efficiency of a downlink cloud radio access network (C-RAN). Here, data is transferred from a baseband unit in the core network to several remote radio heads via a set of edge routers over capacity-limited fronthaul links. The remote radio heads then send the received signals to their users via radio access links. Both data sharing and compression-based strategies are considered for fronthaul data transfer. New mixed-integer nonlinear problems are formulated, in which the ratio of network throughput and total power consumption is maximized. These challenging problem formulations include practical constraints on routing, predefined minimum data rates, fronthaul capacity, and maximum remote radio head transmit power. By employing the successive convex quadratic programming, iterative algorithms are proposed with guaranteed convergence to the Fritz John solutions of the formulated problems. Significantly, each iteration of the proposed algorithms solves only one simple convex program. Numerical examples with practical parameters confirm that the proposed joint optimization designs markedly improve the C-RAN's energy efficiency compared to benchmark schemes. They also show that the fronthaul data-sharing strategy outperforms its compression-based counterpart in terms of energy efficiency, in both single-hop and multi-hop network scenarios.

This paper considers the problem of eigenstructure assignment for output feedback control. We introduce a new method for partial eigenstructure assignment, which allows to place the eigenelements (λi,vi,wi) simultaneously. This is possible by a combination of linear algebra and nonlinear optimization techniques. The advantage of the new approach is illustrated through the control of a launcher in atmospheric flight.