Cell-free networks can effectively reduce interference due to diversity gain.Two key technologies,access point(AP)clustering and transceiver design,play key roles in cell-free networks,and they are implemented at diff...Cell-free networks can effectively reduce interference due to diversity gain.Two key technologies,access point(AP)clustering and transceiver design,play key roles in cell-free networks,and they are implemented at different layers of the air interface.To address the issues and obtain global optimal results,this paper proposes an uplink joint AP clustering and receiver optimization algorithm,where a cross-layer optimization model is built based on graph neural networks(GNNs)with low computational complexity.Experimental results show that the proposed algorithm can activate fewer APs for each user with a small performance loss compared with conventional algorithms.展开更多
Cell-free network is a promising architecture with numerous merits in energy efficiency and macro diversity,which is easy and flexible to integrate with other communication technologies.However,its current network top...Cell-free network is a promising architecture with numerous merits in energy efficiency and macro diversity,which is easy and flexible to integrate with other communication technologies.However,its current network topology where access points(APs)are connected to a central processing unit(CPU)to jointly serve the users,causes huge burden to the fronthaul network.To deal with this problem,in this paper,we first combine thoughts in user-centric(UC)network where users are served by selected subset of APs.Then,we propose a successful transmission probability(STP)based AP clustering scheme to reduce the fronthaul capacity requirement(FCR).By using stochastic geometry and proper approximation methods,the approximated STP calculation expression is derived.Numerical simulations demonstrate that the obtained STP expression can provide a tight approximation compared to Monte Carlo simulation results under different system parameters while keeping the computation tractable.Furthermore,the relationship between the FCR and the STP threshold is formulated as a clustering optimization problem,which gives insights on clustering design in UC-CF network systems.We show by simulation results that the proposed scheme requires less fronthaul capacity than the original CF approach while ensuring the STP performance.展开更多
ultra-Dense Network(UDN)has been envisioned as a promising technology to provide high-quality wireless connectivity in dense urban areas,in which the density of Access Points(APs)is increased up to the point where it ...ultra-Dense Network(UDN)has been envisioned as a promising technology to provide high-quality wireless connectivity in dense urban areas,in which the density of Access Points(APs)is increased up to the point where it is comparable with or surpasses the density of active mobile users.In order to mitigate inter-AP interference and improve spectrum efficiency,APs in UDNs are usually clustered into multiple groups to serve different mobile users,respectively.However,as the number of APs increases,the computational capability within an AP group has become the bottleneck of AP clustering.In this paper,we first propose a novel UDN architecture based on Mobile Edge Computing(MEC),in which each MEC server is associated with a user-centric AP cluster to act as a mobile agent.In addition,in the context of MEC-based UDN,we leverage mobility prediction techniques to achieve a dynamic AP clustering scheme,in which the cluster structure can automatically adapt to the dynamic distribution of user traffic in a specific area.Simulation results show that the proposed scheme can highly increase the average user throughput compared with the baseline algorithm using max-SINR user association and equal bandwidth allocation,while it guarantees at the same time low transmission delay.展开更多
基金supported in part by National Natural Science Foundation of China under Grant No.62171474。
文摘Cell-free networks can effectively reduce interference due to diversity gain.Two key technologies,access point(AP)clustering and transceiver design,play key roles in cell-free networks,and they are implemented at different layers of the air interface.To address the issues and obtain global optimal results,this paper proposes an uplink joint AP clustering and receiver optimization algorithm,where a cross-layer optimization model is built based on graph neural networks(GNNs)with low computational complexity.Experimental results show that the proposed algorithm can activate fewer APs for each user with a small performance loss compared with conventional algorithms.
文摘Cell-free network is a promising architecture with numerous merits in energy efficiency and macro diversity,which is easy and flexible to integrate with other communication technologies.However,its current network topology where access points(APs)are connected to a central processing unit(CPU)to jointly serve the users,causes huge burden to the fronthaul network.To deal with this problem,in this paper,we first combine thoughts in user-centric(UC)network where users are served by selected subset of APs.Then,we propose a successful transmission probability(STP)based AP clustering scheme to reduce the fronthaul capacity requirement(FCR).By using stochastic geometry and proper approximation methods,the approximated STP calculation expression is derived.Numerical simulations demonstrate that the obtained STP expression can provide a tight approximation compared to Monte Carlo simulation results under different system parameters while keeping the computation tractable.Furthermore,the relationship between the FCR and the STP threshold is formulated as a clustering optimization problem,which gives insights on clustering design in UC-CF network systems.We show by simulation results that the proposed scheme requires less fronthaul capacity than the original CF approach while ensuring the STP performance.
基金This work was partially supported by the National Natural Science Foundation of China(61801208,61671233,61931023)the Jiangsu Science Foundation(BK20170650)+2 种基金the Postdoctoral Science Foundation of China(BX201700118,2017M621712)the Jiangsu Postdoctoral Science Foundation(1701118B)the open research fund of National Mobile Communications Research Laboratory(2019D02).
文摘ultra-Dense Network(UDN)has been envisioned as a promising technology to provide high-quality wireless connectivity in dense urban areas,in which the density of Access Points(APs)is increased up to the point where it is comparable with or surpasses the density of active mobile users.In order to mitigate inter-AP interference and improve spectrum efficiency,APs in UDNs are usually clustered into multiple groups to serve different mobile users,respectively.However,as the number of APs increases,the computational capability within an AP group has become the bottleneck of AP clustering.In this paper,we first propose a novel UDN architecture based on Mobile Edge Computing(MEC),in which each MEC server is associated with a user-centric AP cluster to act as a mobile agent.In addition,in the context of MEC-based UDN,we leverage mobility prediction techniques to achieve a dynamic AP clustering scheme,in which the cluster structure can automatically adapt to the dynamic distribution of user traffic in a specific area.Simulation results show that the proposed scheme can highly increase the average user throughput compared with the baseline algorithm using max-SINR user association and equal bandwidth allocation,while it guarantees at the same time low transmission delay.
