In 6G network with extreme MIMO a potential misalignment exists between UE-centric rank selection and the internal link adaptation state of the gNB, specifically the outer loop link adaptation offset. Since this misalignment leads to inefficient resource utilization and degraded quality-of-service, it must be avoided. We present a novel adaptive rank selection algorithm that allows the gNB to dynamically influence the rank selection of the UE to bridge this gap. Our solution incorporates a gNB-configurable rank selection margin, α, which imposes a minimum required throughput gain to justify selecting a higher rank. This margin constrains the greedy rank selection by the UE, and helps aligning it with the link adaptation state of gNB. System-level results confirm that our method outperforms the conventional greedy baseline, boosting XR capacity from 14 to 15 users per cell and delivering consistent UE throughput gains across all measured percentiles for best-effort file transfer traffic.

While extended reality (XR) applications are expected to be one of the cornerstone use cases for the 6G mobile communications system, the actual performance and capacity results for XR in 6G are yet to be rigorously established. This paper bridges this gap by developing a detailed system model for a candidate 6G network and using it for evaluating the performance of XR applications through extensive system-level simulations. We investigate three enablers for the 6G networks: new spectrum utilization, extreme multiple-input multiple-output (MIMO) methods, and advanced radio resource management (RRM) techniques. Our results demonstrate a 3 times XR capacity improvement over 5G baselines at 3.5 GHz using enhanced CSI feedback framework and advanced RRM techniques. Furthermore, by utilizing the new 7 GHz upper-mid band spectrum with extreme MIMO and a larger bandwidth, XR capacity can be further improved by approximately a factor of 10, supporting more than 50 users per cell. These findings provide the first quantitative evidence of 6G’s potential for XR and offer a crucial baseline for future 6G development.

Sub-band full duplex (SBFD) is a novel duplexing scheme that enables concurrent uplink (UL) and downlink (DL) transmission on non-overlapping frequency resources within an unpaired carrier. Earlier studies indicated that to unleash the full potential gain of SBFD, the radio frame configuration must be adapted to traffic load condition and network characteristics. This paper proposes a centralized deep reinforcement learning framework, based on soft actor-critic for discrete action setting (SAC-Discrete), with a modified prioritized replay mechanism, to train a model that dynamically determines the best frame configuration based on cells' buffer statuses. This method aims to maximize the geometric mean of user throughput, ensuring fairness across all users in both link directions. Simulation results demonstrate that our proposed solution improves UL performance by maximizing the geometric mean of user throughput under varying offered loads.

The introduction of sub-band full duplex (SBFD) with simultaneous downlink and uplink transmissions has the potential to improve the uplink performance as compared to traditional static time division duplexing (TDD) with sparse uplink resources. However, SBFD benefits are limited by cross-link interferences caused by the simultaneous uplink and downlink transmissions. This paper proposes to dynamically reduce the 5G base station (gNB) transmission power during SBFD slots to mitigate the gNB self- and gNB-to-gNB-interferences, as well as corresponding link adaptation enhancements to unleash the full potential of such techniques. The SBFD system-level performance is evaluated with different levels of power reduction under varied offered load conditions for urban macro (UMa) and dense urban (DU) environments. Results indicate that for the DU, applying a 10 dB gNB transmit power reduction for SBFD slots leads to a 187% improvement of cell-edge user uplink throughput under high offered load (compared to static TDD). For UMa, at most 4 dB SBFD gNB transmit power reduction is recommended for improving the cell-edge uplink throughput while still ensuring downlink acceptable data rates.

This paper presents a performance evaluation of sub-band full duplex (SBFD) for an indoor office environment. SBFD has the potential to improve packet latency by utilizing non-overlapping uplink and downlink frequency resources simultaneously. To evaluate the SBFD performance, system- level simulations are conducted with different packet sizes and varied offered load conditions. The simulation results reveal a clear dependence of SBFD performance on the packet size of the traffic model. For small packet cases, up to 55% improvement of the average packet latency in the uplink and up to 24% in the downlink is achieved in SBFD, as compared to static TDD, when a 90 dB ratio of self-interference mitigation (RSI) is ensured. For large packet cases, including an uplink-only slot in the SBFD frame structure has the potential to improve the uplink latency, even though this adjustment leads to increased downlink latency. Furthermore, simulation results indicate that the has a noticeable impact on the SBFD performance in the downlink.

The very low block error rate (BLER) targets required in ultra-reliable low-latency communications (URLLC) call for new channel state information (CSI) feedback enhancements for an accurate link adaptation (LA) in the radio interface. This paper describes and analyses two new feedback reporting types, in addition to the legacy positive and negative acknowledgements (ACK/NACK) feedback, based on a form of soft-ACK reports. The aim is to improve the outer loop link adaptation (OLLA) accuracy in 5G New Radio (5G NR) based URLLC wireless communications systems where the amount of NACK events will be negligible. These schemes are based on the physical downlink data channel (PDSCH) decoding performance. In particular, the methods are based on an indication of the decoding margin of a PDSCH transmission, for the so-called Scheme A, and on an indication of the estimated block error probability (BLEP) of a PDSCH transmission, for the so-called Scheme B. In addition, two interference measurement (IM) approaches are analysed and compared based on non zero power CSI reference signal (NZP CSI-RS) and CSI interference measurement (CSI-IM) resources. The results show that Scheme A does not converge to any predefined BLER target while Scheme B allows faster convergence times towards the predefined target compared to legacy OLLA scheme. These results indicate that a new feedback reporting approach based on the estimated BLEP is suitable to achieve the tight latency requirements in URLLC scenarios together with efficient radio link performance.

Due to mobility and interference, using enterprise Wi-Fi for communication in industrial networks can result in control loop latencies exceeding 100 ms for at least 0.1% of the time, even in those cases where Wi-Fi handover-specific parameters have been optimized, making the technology unfit for Industrial IoT (IIoT) with strict communication reliability requirements. To improve its performance, this paper presents a novel approach towards the design and implementation of a radio-aware multi-connectivity concept using a layer-4 scheduling mechanism. Two packet scheduling mechanisms are presented: packet duplication and best path scheduling. A mobility coordinator scheme is used to improve the performance of the packet schedulers by preventing simultaneous handovers and ensures the STAs connect to different APs. By using this multi-connectivity solution, a significant performance improvement was observed, cutting down the latencies of the system to 30-80 ms at the 99.9%-ile of reliability (depending on the operational conditions). Furthermore, by applying the proposed schemes, Wi-Fi handovers delays can be fully mitigated allowing for true seamless roaming in mobile conditions.

5G New Radio (NR) technology is expected to enable the wireless transition of manufacturing processes in modern factories. In order to proper dimension the 5G system, accurate models of the data and control traffic in industrial use cases are needed. In this paper, we analyze the industrial traffic empirically measured in different Danish factories and compare its characteristics to the modeling assumptions used in the 3GPP for 5G NR performance analyses. In particular, three relevant use cases are studied: a unit test cell for actuator calibration, a visual inspection cell, and the control of an autonomous mobile robot (AMR). Our results indicate that some of the relevant 3GPP assumptions for industrial traffic such as exponential packet-interarrival time (PIAT) for aperiodic traffic and fixed packet sizes are only partially corroborated by experimental evidence. Moreover, industrial traffic can exhibit a large burstiness that may lead to conservative admission control and radio resource allocation.

Accurate downlink link adaptation is a major challenge for ultra-reliable and low-latency communications (URLLC) as a consequence of the random and unpredictable load variations at the interfering cells. To address this problem, this paper introduces enhancements to the channel quality indicator (CQI) measurement and reporting procedures for 5G New Radio (NR). The goal is to accurately estimate and report the lower percentiles of the user channel quality distribution. First, a simple and efficient technique is proposed for filtering the channel quality samples collected at the user equipment and, accordingly, estimating tail signal-to-interference-and-noise (SINR) performance. Second, a new CQI reporting format is introduced which better guides downlink scheduling and link adaptation decisions of small URLLC payloads at the gNB. The benefits of the proposed solutions are evaluated via advanced system-level simulations, where it is shown that the proposed solutions significantly outperform existing CQI measurement and reporting schemes. For instance, the 99.999% percentile of the experienced latency is reduced from 1.3 ms to 0.86 ms for the case when URLLC traffic is multiplexed with enhanced mobile broadband (eMBB) traffic.

In this work, we analyze the end-to-end latency measured in a client-server application that emulates the traffic requirements for the Unmanned Aerial Vehicle (UAV)'s Command and Control (C2) link. The connectivity is provided by two real LTE-A networks to a client attached to a flying UAV. Measurements are performed at 4 different heights: ground level, 15 m, 40 m and 100 m. In single operator scenarios, the reliability measured at the target latency, 50 ms, was between 99.6 % and 97.6 % in downlink, and 91.3% and 99.4% in uplink. These results are below the 99.9 % target reliability defined for UAVs and they show that several consecutive packets can be missed when the radio link connectivity degrades, leading to high (> 1 s) values for the 99.9%-ile of latency. To circumvent this, a dualoperator hybrid access scheme is proposed in this paper. The results show that the hybrid access strategy managed to reach the performance requirements in most cases. The solution shows potential to enable C2 over cellular networks, without requiring optimization or modifications in the network.

We address the problem of resource allocation and packet scheduling for a mixture of ultra- reliable low-latency communication (URLLC) and enhanced mobile broadband (eMBB) traffic in a fifth generation New Radio (5G NR) networks. A novel resource allocation method is presented that is latency, control channel, hybrid automatic repeat request (HARQ), and radio channel aware in determining the transmission resources for different users. This is of high importance for the scheduling of URLLC users in order to minimize their latency, avoid unnecessary costly segmentation of URLLC payloads over multiple transmissions, and benefit from radio channel aware multi-user diversity mechanisms. The performance of the proposed algorithm is evaluated with an advanced 5G NR compliant system level simulator with a high degree of realism. Simulation results show promising gains of up to 98% latency improvement for URLLC traffic and 12% eMBB end-user throughput enhancement as compared to conventional proportional fair scheduling.

This paper presents the implementation framework and performance evaluation of a wireless Manufacturing Execution System (MES) for Industry 4.0. The proposed solution is based on self- configuring multi-access gateways that enable seamless transport of delay-tolerant industrial Ethernet control data traffic over LTE or Wi-Fi (or both, using hybrid-access techniques). The wireless MES solution has been deployed at the Smart Production Lab facilities at Aalborg University, allowing the removal of Ethernet cables between modules in a production line setup, and thus enabling a faster re-configuration of the production facilities. The performance of the wireless MES solution is benchmarked against the reference Ethernet case in terms of latency and packet loss. The trials revealed that, despite the increase in latency and packet loss as compared to the reference operation over Ethernet, both LTE and Wi-Fi under different conditions were able to reliably support the production control operations at MES level without discernable difference on the overall functionality of the system.

In this paper, a new multi-connectivity protocol is introduced for mission critical applications that use wireless 5G multi-UE devices. It extends the MP-QUIC transport protocol by adding selective data duplication with strict prioritization to support a mix of critical and non-critical applications on a single end-to-end connection; e.g. suitable for self-driving cars getting mission critical drive control commands from the network (priority) while receiving map- updates (background). The Selective Redundant MP- QUIC solution simultaneously provides bandwidth aggregation and higher speeds for background data while ensuring maximum reliability and availability for the priority data by duplicating it on the available paths. The proposed algorithm is first evaluated using an emulated fixed-delay channel as well as a well-known Internet connection model. Next, a highway drive test is conducted using two LTE modems, resembling a typical autonomous car scenario. The tests show that the algorithm can protect the critical data flow also in presence of a background load of 2 Mb/s and across a wide range of channel conditions. Using our proposed MP-QUIC solution over two LTE connections, the 99.9%-tile of critical traffic latency shows a 5 times improvement versus using a single-path connection and 3 times versus using state of the art MP-QUIC protocols.

In this paper we study the potential of commercial Long Term Evolution (LTE) networks for supporting reliable low-latency applications. To this end, we use one-way end-to-end latency measurements of a commercial LTE network deployed in Vodafone R&D's Innovation Lab where we control the over-the-air traffic and air interface configuration. The experimental results show that the achievable latency and reliability is mainly limited by two features of the LTE radio system: the configured Discontinuous Reception (DRX) pattern and the Scheduling Request (SR) periodicity of the UE. The one-way latencies vary between 40 and 95 ms at the 99.9% percentile, depending on the inter-arrival time of the generated traffic. By disabling or applying less-restrictive DRX and SR patterns and using an aggressive uplink allocation technique, we show how the one-way latency performance improves significantly, e.g. as low as 7 ms with 99.9% probability regardless of the traffic characteristics. For cases where configuring the radio interface is not feasible, we show different options available at the application layer for improving the latency performance of the application. For instance, the application may frequently transmit small packets to reduce the probability of the UE going to DRX mode, or coordinate the message generation interval with the SR opportunities and DRX patterns of the underlying radio system.

Autonomous vehicle applications require sub-100 ms message latency with a high success probability. While this will be supported by upcoming fifth generation networks, hybrid access technology using multiple LTE connections, may pave the way for the autonomous vehicle applications today. In current LTE networks, handovers often lead to long data interruption. Therefore, we study the handovers and the related data interruption, and how performance can be improved through hybrid access, using redundant transmissions over multiple connections. The study is based on drive tests, where four QualiPoc measurement smartphones are connected to different LTE networks simultaneously, performing synchronized ping latency measurements, and recording handover events. The results show that a handover event during a ping measurement, prolongs the average latency from 60-80 ms to more than 200 ms. However, the handovers do not occur simultaneously in the measured LTE networks. Therefore, hybrid access with two connections is shown to reduce the handover outage by a factor 60. Furthermore, the 99.9 %-tile latency is reduced 66 %, by using two simultaneous connections, as compared to the best single network measurement.

In this paper, we present an exhaustive system level analysis of using preemptive scheduling for latency critical traffic in coexistence with mobile broadband for the 3GPP 5G New Radio. Enhanced recovery and HARQ retransmission mechanisms exploiting base station a priori knowledge of punctured radio resources from using preemptive scheduling are proposed. It is demonstrated that a scheme with HARQ multi-bit feedback and selective retransmission of punctured resources is an attractive solution. Furthermore, the performance sensitivity from using either fully interleaved or frequency-first code block layouts is assessed. The impact on the mobile broadband performance is evaluated at TCP-level, studying both the penalty on throughput and smoothened TCP round trip time to assess how preemptive scheduling affects the end-to-end performance of other traffic.

This paper presents solutions for efficient multiplexing of ultra-reliable low-latency communications (URLLC) and enhanced mobile broadband (eMBB) traffic on a shared downlink channel. Such a scenario presents multiple challenges in the context of radio resource scheduling, link adaptation and inter-cell interference, which are identified and addressed throughout the paper. Specifically, a joint link adaptation and resource allocation technique is proposed that dynamically adjusts the block error probability (BLEP) of the URLLC payload transmissions in accordance with the instantaneous experienced load per cell. Extensive system-level simulations of the downlink performance show promising gains of this technique, reducing the URLLC latency from 1.3 ms to 1 ms at the 99.999% percentile, with less than 10% degradation of the eMBB throughput performance as compared to conventional scheduling policies.

In this paper, we present a punctured scheduling scheme for efficient transmission of low latency communication (LLC) traffic, multiplexed on a downlink shared channel with enhanced mobile broadband traffic (eMBB). Puncturing allows to schedule eMBB traffic on all shared channel resources, without prior reservation of transmission resources for sporadically arriving LLC traffic. When LLC traffic arrives, it is immediately scheduled with a short transmission by puncturing part of the ongoing eMBB transmissions. To have this working efficiently, we propose recovery mechanisms for punctured eMBB transmissions, and a service-specific scheduling policy and link adaptation. Among others, we find that it is advantageous to include an element of eMBB-awareness for the scheduling decisions of the LLC transmissions (i.e. those that puncture ongoing eMBB transmissions), to primarily puncture eMBB transmission(s) that are transmitted with low modulation and coding scheme index. System level simulations are presented to demonstrate the benefits of the proposed solution.

In this paper, we investigate the operation of eICIC in a realistic deployment based on site specific data from a dense urban European capital area. Rather than the classical semi-static and common network-wide configuration, the importance of having highly dynamic and distributed mechanisms that are able to adapt to local environment conditions is revealed. We propose a promising opportunistic cell association algorithm and a generalized method for fast muting adaptation. The performance results show that the traditional semi-static eICIC configuration leads to modest gains in realistic deployments, whereas the set of proposed fast dynamic eICIC algorithms leads to capacity gains on the order of 35-120% depending on the local environment characteristics. In the analysis of the performance results for the site specific use case, it is furthermore highlighted how those deviate from typical findings from 3GPP standardized HetNet scenarios.

Dual Connectivity (DC) has been studied and proved to be an effective solution to deal with the fragmented resources in deployment scenarios where the macro and small cells use different frequency carriers. The performance of DC has mainly been analyzed using generic network models such as those proposed by the 3GPP. However, the benefits of DC in real networks have not been proved. In this paper, we investigate the performance of DC in a realistic deployment from a big European city. Additionally, a novel opportunistic cell selection technique is also proposed. Results show that DC does improve the performance in this realistic layout. Due to the uneven load distribution observed in realistic deployments, DC is able to provide fast load balancing gains also at relatively high load - and not only at low load as typically observed in 3GPP scenarios. For the same reason, the proposed cell selection technique that aims at performing intra-layer load balancing shows promising benefits in this irregular deployment.

The support of mission-critical communication (MCC) opens the possibility to implement a broad range of novel applications. V2X communication for traffic safety and automation is, among others, one of these innovative applications expected to bring big benefits to society: accidents are prevented, driving times are reduced, and carbon dioxide is saved. In this regard, we first present a system model and fundamental definitions of reliability, latency and availability. Relying on these definitions, a systematic review of requirements for the huge variety of V2X applications is provided, including insights into the expected evolution towards autonomous driving. The many challenges introduced by V2X use cases are emphasized and compared to today's wireless system capabilities. Finally, we give our vision on the design of future radio technologies for the support of this kind of communications.

The arrival of mission critical applications in the context of vehicular, medical and industrial wireless communications calls for reliability constraints never seen before in cellular systems. Enhanced Inter-Cell Interference Coordination (eICIC) has been widely investigated in the context of LTE-A Heterogeneous Networks, but always with load balancing and resource partitioning purposes. Given the broad range of new use cases targeting ultra high reliability, we propose the use of on-demand eICIC for reducing the BLER of the retransmissions of critical users while minimizing the impact to the rest of the network. Combined with a ROot Cause Aware HARQ (ROCA-HARQ), which provides additional information when a transmission fails, the joint mechanism is relevant for any LTE/LTE-A deployment and can be easily implemented in a real network. System-level simulations show attractive BLER reductions up to 80% with little impact in throughput performance (loss in user throughput below 6%).

Dual connectivity (DC) is a promising technique to boost the user throughput performance by allowing user equipments (UEs) to receive data simultaneously from a macro cell and a small cell. In order to ensure high degree of realism and practical relevance of the results, we investigate the performance of DC in a realistic deployment based on three-dimensional data from a dense urban European capital area, assuming realistic flow control on the backhaul connections between the macro and small cell eNBs. It is found that the configuration of UEs with DC plays a critical role in the performance of DC under realistic conditions. A modified opportunistic cell association algorithm is proposed. Simulation results show that with proper configuration of UEs with DC, the performance of DC exhibits similar gains as observed in generic 3GPP scenarios.