Federated learning is an effective method to train a machine learning model without requiring to aggregate the potentially sensitive data of agents in a central server. However, the limited communication bandwidth, the hardware of the agents and a potential application-specific latency requirement impact how many and which agents can participate in the learning process at each communication round. In this paper, we propose a selection metric characterizing each agent’s importance with respect to both the learning process and the resource efficiency of its wireless communication channel. Leveraging this importance metric, we formulate a general agent selection optimization problem, which can be adapted to different environments with latency or resource-oriented constraints. Considering an example wireless environment with latency constraints, the agent selection problem reduces to the 0/1 Knapsack problem, which we solve with a fully polynomial approximation. We then evaluate the agent selection policy in different scenarios, using extensive simulations for an example task of object classification of European traffic signs. The results indicate that agent selection policies which consider both learning and channel aspects provide benefits in terms of the attainable global model accuracy and/or the time needed to achieve a targeted accuracy level. However, in scenarios where agents have a limited number of data samples or where the latency requirement is very stringent, a pure learning-based agent selection policy is shown to be more beneficial during the early or late stages of the learning process.@en

—Wireless communication networks provide a critical infrastructure, particularly in emergency situations due to disruptive events such as natural disasters or terrorist attacks. However, in these kinds of scenarios part of the network may no longer be operational and a traffic hotspot may emerge, which may result in coverage and/or capacity issues. Deploying self-steering drone-mounted base stations offers a potential method to quickly restore coverage and/or provide capacity relief in such situations, but appropriate positioning is crucial in order for a drone base station to be truly effective. Motivated by that challenge, we propose a data-driven algorithm to optimize the position of a drone base station in a scenario with a site failure and emergence of a traffic hotspot. We demonstrate that the use of a drone, when properly positioned, yields significant performance gains, and that our algorithm outperforms benchmark mechanisms in a wide range of scenarios. In addition, we show that our algorithm is able to find a near-optimal position for the drone in a reasonable amount of time, and even has the ability to track the optimal position in case of a moving hotspot.

@en

Reliable mobile communications is of critical importance, and should be maintained even in case of extremely crowded events or emergency scenarios. In such scenarios the deployment of drone-mounted base stations offers an agile and cost-efficient way to sustain coverage and/or provide capacity relief. In this paper we develop an analytical method to estimate the blocking and coverage probabilities of drone-assisted cellular networks using information that is readily available from network planning tools. We demonstrate how this method can be used to determine the minimum required number of drones and their corresponding locations for a given target performance level.@en

Social VR aims at enabling people located at different places to communicate and interact with each other in a natural way. It poses extremely strong throughput and latency requirements on the underlying communication networks. This paper investigates the potential of using cross-layer design approaches for radio access scheduling in order to realize these challenging requirements in (beyond) 5G networks. In particular, we provide an in-depth simulation study of the performance/capacity gains that can be achieved by exploiting the end-to-end latency budget and/or video frame type as cross-layer information in the scheduling decisions, and show how the benefits depend on the actual social VR scenario. This study further reveals the importance of using application-level metrics such as PSNR or SSIM rather than traditional network-level metrics like the packet drop rate in the performance assessment.@en

In a 5G Radio Access Network (RAN), different features are offered as solutions to serve traffic with diverse characteristics and requirements, including flexible numerology, (non-)pre-emptive mini-slot based scheduling and network slicing. In this paper, we present an extensive simulation-based assessment of the relative merit of these distinct 5G features in the context of a smart city environment. We further derive the optimal feature combination and associated configuration which best handles the services related to the smart city environment given their performance requirements. The obtained insights confirm the commonly argued potential of slicing, emphasizing that the optimal configuration of the slice-specific numerology depends not only on the nature of the handled services but also on the selected RAN features. Among these features, non-preemptive mini-slot based scheduling and idle resource sharing reveal significant performance potential.@en

In this paper, we focus on the link density in random geometric graphs (RGGs) with a distance-based connection function. After deriving the link density in D dimensions, we focus on the two-dimensional (2D) and three-dimensional (3D) space and show that the link density is accurately approximated by the Fréchet distribution, for any rectangular space. We derive expressions, in terms of the link density, for the minimum number of nodes needed in the 2D and 3D spaces to ensure network connectivity. These results provide first-order estimates for, e.g., a swarm of drones to provide coverage in a disaster or crowded area.

@en

One of the most challenging applications targeted by evolving (beyond-)5G technology is virtual reality (VR). Particularly, 'Social VR' applications provide a fully immersive experience and sense of togetherness to users residing at different locations. To support such applications the network must deal with huge traffic demands, while keeping end-to-end latencies low. Moreover, the radio access network must deal with the volatility and vulnerability of mmWave radio channels, where even small movements of the users may have substantial effects on the Quality of Experience. We present an integral modelling framework for feasibility assessment and performance optimization of the radio access network for Social VR applications in indoor office scenarios. Using the presented modelling approach, we conduct an extensive simulation-based assessment to determine the performance impact of head motion, the frequency band (3.5 GHz, 26 GHz) and radio network configurations, and derive the required carrier bandwidth for a range of 'Social VR' scenarios. Insights into these issues are a prerequisite for setting up guidelines for network deployment and configuration as well as for the development of (AI/ML-based) methods for dynamic resource management to optimally support Social VR applications. @en

With increasing network complexity, intelligent mechanisms to efficiently achieve the required quality of service of wireless-enabled applications are being developed, especially for industrial environments due to the onset of the fourth industrial revolution. In this paper, the potential benefits of wireless channel quality prediction for two of the three major use cases supported by 5G viz. enhanced Mobile BroadBand (eMBB) and Ultra-Reliable Low Latency Communication (URLLC) are quantified in an industrial indoor environment through simulations. Our analysis shows that the ability to perform perfect prediction improves the 10th user throughput percentile by up to 125% for eMBB use case and decreases the 90th resource utilization percentile by up to 37% for URLLC use case. Furthermore, the maximum tolerable prediction inaccuracy is found to be up to 5 dB and 0.35 dB for eMBB and URLLC use cases, respectively.

@en

Drone base stations can help safeguard coverage and provide capacity relief when cellular networks are under stress. Examples of such stress scenarios are events with massive crowds or network outages. In this paper we focus on a disaster scenario with emergence of a traffic hotspot, where agile drone positioning and load management is a critical issue. In order to address this challenge, we propose and assess a data-driven algorithm which leverages real-time measurements to dynamically optimize the 3D position of the drone as well as a cell selection bias tuned for optimized load management. We compare the performance with three benchmark scenarios: i) no drone; ii) a drone positioned above the failing site; and iii) a drone with a statically optimized position and cell selection bias. The results demonstrate that the proposed algorithm significantly improves the call success rate and achieves close to optimal performance. @en

Network slicing has been introduced in 5G networks as an enabling feature for the effective Quality of Service (QoS) provisioning to multiple service classes with distinct performance requirements. When applied in the Radio Access Network (RAN), a class-specific slice is assigned a set of radio resources and can furthermore be optimally configured in terms of the applied numerology and packet scheduler. As both the optimal numerology and the most suitable packet scheduler may be different for e.g. a class of Latency-Constrained (LC) and a class of Throughput-Oriented (TO) services, the potential of slicing is clear. However, the inherent trunking loss incurred when applying slicing with dedicated resources provides an argument against such slicing. In this paper we demonstrate that the performance and traffic handling capacity in an optimally configured non-sliced scenario may exceed that attained when using segregated individually optimised slices. To that end, we use simulations to assess the best-performing numerology and packet scheduler for a sliced scenario with LC and TO services. We then compare the thus optimised sliced scenario with an optimal non-sliced scenario and show that the non-sliced scenario can serve about 20% more traffic than the sliced scenario while satisfying the same class-specific QoS requirements.@en

The use of drone base stations offers an agile mechanism to safeguard coverage and provide capacity relief when cellular networks are under stress. Such stress conditions can occur for example in case of special events with massive crowds or network outages. In this paper we focus on a disaster scenario with emergence of a hotspot, and analyze the impact of the drone position (altitude, horizontal position) and selection bias on the network performance. We determine the optimal settings of these control parameters as a function of the hotspot location, and demonstrate that the optimized values can drastically reduce the fraction of failed calls.

@en

With the recent adoption of millimeter-wave spectrum in cellular communications, deployment of active antenna arrays and use of beamforming become vital to compensate for the increased path loss. However, directional high-frequency signals may suffer heavy attenuations due to blockage effects. Therefore, blockage modelling that adequately incorporates the effects of beamforming becomes increasingly relevant. We propose a Four Knife-Edge Diffraction with antenna Gain (4KED-G) model, a deterministic approach to model blockage with broad applicability. The 4KED-G model advances upon the existing models in its inclusion of both angular antenna gains and the diffraction from all the four edges of a rectangular screen blocker, leading to a more general and flexible blockage modelling approach compared to existing widely accepted blockage models. We theoretically show that the proposed generalised model incorporates the strengths of these existing models, while overcoming their shortcomings in establishing applicability to wider range of blockage scenarios. We validate the generalised model against known knife-edge diffraction blockage models for specific scenarios.

@en

This study proposes a two-step ML-based multislice radio resource allocation framework for 5G networks, specifically for emergency scenarios and featuring a good tradeoff between complexity and performance. In the first step, call-level resource demands are predicted using supervised ML, which are then aggregated to predict slice-specific resource demands. An innovative method is included in this step to ensure the collection of representative training data for the supervised ML. In the second step, a contextual multi-armed bandit reinforcement learning model is applied to derive the resource allocation among the slices based on the slice-specific resource demand predictions. The simulation results show that the proposed framework outperforms alternative solutions in the defined utility values for priority emergency traffic at the cost of modest performance sacrifice of the background traffic.

@en

We present a simulation-based assessment of the performance potential of distributed MIMO (D-MIMO), multi-user MIMO (MU-MIMO) and particularly the combined D/MU-MIMO operation, for which we extend previously published scheduling and beamforming principles. The assessment study reveals that, when optimizing average user throughput performance, D-MIMO, while fruitless when used in isolation, is very effective when intelligently combined with MU-MIMO. Alternatively, when optimizing the cell edge performance, MU-MIMO, while also shown to be ineffective when used in isolation, is in fact very valuable when accompanied by a suitable configuration of D-MIMO. As an illustrative example, when using a jointly optimised configuration of D/MU-MIMO in a highly loaded urban deployment scenario, a 121% cell edge performance gain can be attained over a scenario using only D-MIMO, and even a demonstrated 153% gain over a scenario where only the MU-MIMO feature is available.

@en

We design and assess novel random access (RA) procedures for use in 5G radio access networks in support of machine-type applications requiring Ultra-Reliable Low-Latency Communications (URLLC). In a simulation-based performance assessment the proposed RA procedures are compared against LTE Cat-M1 (or LTE-M), the current LTE-based access technology for such applications, considering a Factory of the Future scenario and a realistic density and mix of URLLC and non-URLLC devices. As a key outcome, the most promising Two-step RA with Feedback procedure is demonstrated to be able to satisfy the 3GPP-imposed 10 ms (50 ms) latency requirement of URLLC (non-URLLC) applications in 99.901% (99.997%) of the cases, with a theoretical minimum access latency of only 5 ms.

@en

A suitability assessment and performance optimisation is presented of narrowband Internet of Things (NB-IoT) cellular technology for use in smart energy distribution networks. The focus is on the reliable and timely delivery of outage restoration and management (ORM) messages at the event of a local or regional power outage. Both the cellular NB-IoT and the energy distribution networks are modelled in a system-level simulator, which is used to carry out an extensive sensitivity analysis of the ORM service performance w.r.t. various radio network configurations in different environments. In particular, different packet schedulers are proposed and analysed, addressing device prioritization and subcarrier allocation as essential mechanisms in optimizing the service performance. Furthermore, we consider all three possible NB-IoT spectral deployment modes: in-band, guard-band and stand-alone deployment. Results show that, with a proposed near-optimal radio network configuration, the reliability of the ORM message delivery is close to 100% for the majority of power outage scenarios, while the observed 95th transfer delay percentile for the ORM messages is within the acceptable limit of 20 s. The study concludes that indeed NB-IoT is a suitable technology for supporting ORM services in smart energy distribution networks.

@en

In this paper, we assess the suitability of NB-IoT (Narrowband Internet of Things) cellular technology for smart grid applications, concentrating on the reliable and timely delivery of Outage Restoration Management (ORM) messages at the event of a local or regional power outage. Using system-level simulations modelling of both the cellular NB-IoT and the energy distribution networks for different environments, we present an extensive sensitivity analysis of the ORM service performance w.r.t. various radio network configurations. In particular, we propose and analyze different packet schedulers, an essential mechanism in optimizing the service performance. A key outcome of the study is the conclusion that indeed NB-IoT is a suitable technology for supporting ORM services in smart grids, accompanied with a proposed near-optimal radio network configuration to best do so.

@en