Sensor-aided wireless edge applications, such as LiDAR-based beam prediction for millimeter-wave communications, demand intelligent on-device processing of high-volume sensor data. However, the computational cost of machine learning models often exceeds the tight power and resource constraints of edge hardware. While quantized neural networks (QNNs) reduce resource requirements, typical FPGA accelerators still rely on power-hungry digital signal processing (DSP) slices and incur avoidable data-movement overheads. To bridge this gap, we propose ArchE-Q, a dataflow accelerator for QNNs combined with efficient data preprocessing. Our design is fundamentally multiplier-less, utilizing: (1) an application-specific first-layer kernel that exploits binarized sensor inputs to remove multipliers; (2) the eXtended Vector Activation Unit (XVAU), fusing convolution, activation, and pooling to reduce buffering and data transfers; and (3) memory-centric buffering for efficient data reuse. Implemented on a Xilinx ZCU104 FPGA, ArchE-Q achieves 13.3% lower latency and up to 28% lower dynamic power than the FINN-R baseline, while eliminating DSP usage.

Millimeter-wave (mm-wave) and sub-terahertz communication are cornerstones for 6G networks, but are hindered by their reliance on highly directional beams and the high latency of conventional beam alignment and beam management procedures. Sensor- and machine learning (ML)-based beam prediction offers a low-latency alternative, but remains constrained by hardware limitations at the wireless edge. This paper presents a systematic, hardware-aware analysis of ML models for sensor-aided beam prediction, specifically focusing on LiDAR, spanning deep learning, classical ML, and a novel class of proposed hybrid models that combine a CNN-based feature extractor with lightweight classical classifiers. Using the public DeepSense 6G dataset, models are benchmarked on accuracy, inference latency, memory footprint, and computational complexity. Our results show that classical ML models, particularly linear and tree-based methods, deliver strong accuracy-efficiency trade-offs, achieving up to 95× lower inference latency and 67× smaller memory footprint than CNNs with minimal accuracy loss. The proposed hybrid models further enhance this trade-off, maintaining CNN-level top-5 accuracy while drastically reducing re-training time by up to 20× and enabling fast, environment-specific adaptation. These findings establish classical and hybrid models, supported by efficient preprocessing, as promising candidates for real-time, resource-efficient beam prediction in future 6G edge systems.

Large antenna arrays and beamforming techniques are key components for exploiting the spectrum-rich FR2 bands in next-generation mobile communication networks. Given the site-specific spatio-temporal variations of the mm-wave channel, non-RF sensor inputs and environment awareness can be leveraged to greatly enhance beam management decisions, e.g. via machine learning (ML) techniques. However, the current literature lacks open platforms to gather datasets for the training of such ML techniques and to evaluate novel beam management approaches in real-time, real-world scenarios and full-stack endto-end networks. In this work, we present our SDR-based experimental platform based on OpenAirInterface and are the first to integrate popular low-cost antenna array transceivers, beam sweeping capabilities, and a highly-modular sensor framework and associated interfaces into such a full-stack experimental platform. This enables beam management experimentation in real-world, real-time scenarios and facilitates gathering datasets necessary for developing ML-based beam management protocols that incorporate environment awareness via sensor modalities.

Millimeter-wave (mm-wave) communications require beamforming and consequent precise beam alignment between the gNodeB (gNB) and the user equipment (UE) to overcome high propagation losses. This beam alignment needs to be constantly updated for different UE locations based on beamsweeping radio frequency measurements, leading to significant beam management overhead. One potential solution involves using machine learning (ML) beam prediction algorithms that leverage UE position information to select the serving beam without the overhead of beam sweeping. However, the highly site-specific nature of mm-wave propagation means that ML models require training from scratch for each scenario, which is inefficient in practice. In this paper, we propose a robust cross-environment transfer learning solution for location-aided beam prediction, whereby the ML model trained on a reference gNB is transferred to a target gNB by fine-tuning with a limited dataset. Extensive simulation results based on ray-tracing in two urban environments show the effectiveness of our solution for both inter- and intra-city model transfer. Our results show that by training the model on a reference gNB and transferring the model by fine-tuning with only 5 % of the target gNB dataset, we can achieve 80 % accuracy in predicting the best beam for the target gNB. Importantly, our approach improves the poor generalization accuracy of transferring the model to new environments without fine-tuning by around 75 percentage points. This demonstrates that transfer learning enables high prediction accuracy while reducing the computational and training dataset collection burden of ML-based beam prediction, making it practical for 5G-and-beyond deployments.

Communication at millimeter-wave (mm-wave) enables unprecedented data rates for 5G-and-beyond cellular networks but relies on precise beam alignment to overcome the challenging propagation characteristics at higher frequencies. Leveraging additional sensor-aided context information for beam management, such as location data, has been proposed to reduce the search space of feasible directional beam-pair-links and reduce signaling overhead. So far, evaluation of such location-aided schemes has largely been based on simplified and limited-scope simulation scenarios, without accounting for the complexity and performance implications in real end-to-end cellular networks. In this work, we present REMLAB, a first-of-its-kind framework for REM-based location-aided beam management in the full-stack 5G-NR-compliant network simulator ns-3 and make our code publicly available. Our framework provides access to network performance metrics and parameters that can only be obtained in realistic end-to-end network simulation, such as the signaling overhead required for beam management. We evaluate REMLAB in an example urban mobility scenario and show an improvement in average IP throughput of at least 24% and up to 46% over default 5G-NR by eliminating the need for periodic control signaling. Furthermore, towards the evaluation of practical REM-based schemes, we implement a realistic location error model and show that localization errors increase the radio link failure duration. This highlights that the real-world feasibility of context-based beam management in 5G-and-beyond networks depends on the development and detailed evaluation of sensor-error-aware schemes.

5G-NR in FR2 relies on beamforming and accurate beam alignment between the gNodeB (gNB) and the user equipment (UE) to address challenging channel conditions. Maintaining beam alignment for a mobile UE requires frequent updates via beam-sweeping and directional radio frequency measurements, entailing high delay and signaling overhead. Location-aided beam management has emerged as a promising solution for reducing this overhead by combining UE location with historical directional beam training information to construct radio environment maps (REMs) for effective beam alignment. However, the highly site-specific nature of millimeter-wave (mm-wave) propagation makes REM-based beam management vulnerable to performance degradation due to errors in UE location information, e.g., obtained via global navigation satellite system (GNSS), affecting the accuracy of both REM construction and querying. This paper proposes a 5G-NR standard-compliant solution to mitigate GNSS location error in REM-based beam management, by constructing our rank-REM from multiple UE directional channel measurements and multi-beam monitoring for connected UEs. Our results, based on ray-tracing in an urban environment, show that our proposed REM construction and querying strategy effectively mitigates realistic GNSS location error, substantially enhancing link stability for a mobile UE by reducing radio link failures (RLFs) and handovers, thus enabling effective REM-based beam management for 5G-and-beyond.

While UE-centric cell-free massive MIMO (CF-mMIMO) provides high and uniform throughput performance under the assumption of a uniform propagation environment modeled by the log-distance path loss channel model, the performance under a realistic urban propagation environment is not yet fully addressed. In this paper we conduct the first comparative performance study of CF-mMIMO under both the widely assumed log-distance channel model and the realistic urban propagation environment obtained via raytracing using real 3D city layouts and practical AP locations. Our results show that with the raytracing channel model, CF-mMIMO cannot achieve as high and uniform throughput performance as observed with the log-distance channel model, putting into question the attractiveness in practice of CF-mMIMO for real urban depolyments.

Hybrid beamforming (HBF) multi-user multiple-input multiple-output (MU-MIMO) is a key technology for unlocking the directional millimeter-wave (mm-wave) nature for spatial multiplexing beyond current codebook-based 5G-NR networks. In order to suppress co-scheduled users' interference, HBF MU-MIMO is predicated on having sufficient radio frequency chains and accurate channel state information (CSI), which can otherwise lead to performance losses due to imperfect interference cancellation. In this work, we propose IABA, a 5G-NR standard-compliant beam pair link (BPL) allocation scheme for mitigating spatial interference in practical HBF MU-MIMO networks. IABA solves the network sum throughput optimization via either a distributed or a centralized BPL allocation using dedicated CSI reference signals for candidate BPL monitoring. We present a comprehensive study of practical multi-cell mm-wave networks and demonstrate that HBF MU-MIMO without interference-aware BPL allocation experiences strong residual interference which limits the achievable network performance. Our results show that IABA offers significant performance gains over the default interferenceagnostic 5G-NR BPL allocation, and even allows HBF MU-MIMO to outperform the fully digital MU-MIMO baseline, by facilitating allocation of secondary BPLs other than the strongest BPL found during initial access. We further demonstrate the scalability of IABA with increased gNB antennas and densification for beyond-5G mm-wave networks.

The spectrum-rich millimeter-wave (mm-wave) bands and exploiting multi-antenna technologies are envisioned as a key enabler for future high speed communication networks. 5G-NR in automation and industry requires supporting not only eMBB but also URLLC applications to meet the demands in production processes. Providing high-rate mm-wave coverage in real-world industrial environments is challenging and necessitates detailed and site-specific characterization of the directional link opportunities and beam management requirements for network planning of mm-wave factory deployments. In this paper, we present the results of our large-scale mm-wave measurement study using phased antenna arrays in a machine production hall. We systematically collect received signal strength data over fine-grained 3D TX/RX orientations for 31 spatially-dense RX positions in three typical factory scenarios in the 28 GHz and 60 GHz bands. We study the impact of transmitter placement and operating frequency band on the achievable data rate and the beam management effort considering the data rate demands of next-generation industrial networks. Our results show that the 28 GHz band provides sufficient connectivity to deliver data rates of up to 1 Gbps without the need for sophisticated beam management, which is in strong contrast to outdoor mobile mm-wave applications where active beam tracking is crucial to provide seamless connectivity.

Agile beam management is key to seamless high-speed mm-wave connectivity in the beyond-5G era, given the site-specific spatio-temporal variations of the mm-wave channel. Leveraging non-RF sensor inputs for environment awareness, e.g. via ML techniques, can greatly enhance RF-based beam management. To address the lack of diverse publicly available multi-modal mm-wave datasets for the design of novel beam management approaches and to enable their real-world, real-time evaluation, we present our SDR-based multi-band mm-wave experimental platform which integrates multi-modal sensors towards environment-aware beam management.

While UE-centric cell-free massive MIMO (CF -mMIMO) provides superior throughput performance against traditional network-centric distributed MIMO in static networks, it loses its advantage under mobility due to the effects of channel aging and hand over delay overheads. To preserve the ability of UE-centric CF-mMIMO to offer uniformly high network-wide throughput under mobility, we propose a socially -differentiated hand over scheme. Our scheme differentiates the handover policy for different UEs by their performance level relative to the rest of the network, thus reducing the handover rate while maintaining the necessary handovers for the most “in-need” UEs. Our results show that our scheme achieves high mobility-aware throughput performance for UE-centric CF-mMIMO and significantly outperforms both the existing handover schemes and network-centric MIMO, under both channel aging and handover delay cost.

While cell-free massive MIMO (CF-mMIMO) provides superior throughput performance against traditional cellular networks in static scenarios, the impact of changing mobile channels and serving access point (AP) sets is not yet fully addressed. In this paper we propose an analysis framework for mobile scalable CF-mMIMO taking into account both the channel aging effect and handover cost, by deriving the through-put model of different AP selection methods as functions of UE mobility and AP serving set size. We then use this framework to compare the performance of different AP selection methods for scalable CF-mMIMO, and thereby provide guidelines for future practical mobile network design. Our results show that scalable CF-mMIMO with UE-centric AP selection methods can be superior to the network-centric method only at very low mobility.

Exploiting millimeter-wave (mm-wave) for high-capacity multi-user networks is predicated on jointly performing beam management for seamless connectivity and efficient resource sharing among all users. Beam management in 5G-NR actively monitors candidate beam pair links (BPLs) on the serving cell to simply select the user’s best beam, but neglects the multi-user resource sharing problem, potentially leading to severe throughput degradation on overloaded cells. We propose flexRLM, a coordinator-based flexible radio link monitoring (RLM) framework for multi-user downlink mm-wave networks. flexRLM enables flexible configuration of monitored BPLs on the serving and other candidate cells and beam selection jointly considering link quality and resource sharing. flexRLM is fully 5G-NR-compliant and uses the LTE coordinator in non-standalone mode to continuously update the monitored BPLs via measurement reports from periodic downlink control synchronization signals. We implement flexRLM in ns-3 and present full-stack simulations to demonstrate the superior performance of flexRLM over default 5G-NR RLM in multi-user networks. Our results show that flexRLM’s continuous updating of monitored BPLs improves both link quality and stability. By monitoring BPLs on candidate cells other than the serving one, flexRLM also significantly decreases handover decision delays. Importantly, flexRLM’s low-complexity coordinated load-balancing achieves a per-user throughput close to the single-user baseline.

Scalable cell-free massive MIMO (CF-mMIMO) offers great advantages in network throughput compared to traditional cellular networks by leveraging sophisticated signal processing for interference management. However, this entails cooperation among CPUs controlling serving AP clusters, resulting in a high signalling and computational burden, and in turn, increased energy and economic network costs. It is thus not yet clear whether CF-mMIMO is overall an attractive architecture for future networks. To address this question, we present the first comprehensive energy and economic efficiency analysis of scalable CF-mMIMO with different cluster cooperation levels, where we explicitly quantify the power consumption cost of signalling and computation. Considering the trade-off between high throughput and low energy and economic cost, our results show that scalable CF-mMIMO is an attractive architecture for sustainable "green" future networks but only at low cooperation levels.

We study the problem of joint user association and beam pair link (BPL) allocation in millimeter-wave (mm-wave) cellular networks. We propose two interference-aware strategies – a centralized and a distributed one – and evaluate their performance based on site-specific directional channel data and realistic antenna models. Our results show that using idealized sectored antenna models severely underestimates the spatial interference, considering the non-negligible sidelobes of realistic antenna arrays which strongly limit the achievable spatial separation of the allocated BPLs in mm-wave networks using beam codebooks. We also show that intra-cell interference is the dominant interference component for all allocated users, in contrast to assumptions in the prior literature. By exploiting non line-of-sight BPLs, our interference-aware strategies achieve significant performance gains over interference-agnostic 5G-NR default user association to the strongest base station and BPL, as well as outperforming a centralized, load-balancing literature benchmark. Our proposed strategies rely solely on downlink 5G-NR reference signals for channel state information updates, making them attractive for practical codebook-based mm-wave cellular networks.

While scalable cell-free massive MIMO (CF-mMIMO) shows advantages in static conditions, the impact of its changing serving access point (AP) set in a mobile network is not yet addressed. In this paper we first derive the CPU cluster and AP handover rates of scalable CF-mMIMO as exact numerical results and tight closed form approximations. We then use our closed form handover rate result to analyse the mobility-aware throughput. We compare the mobility-aware spectral efficiency (SE) of scalable CF-mMIMO against distributed MIMO with pure network- and UE-centric AP selection, for different AP densities and handover delays. Our results reveal an important trade-off for future dense networks with low control delay: under moderate to high mobility, scalable CF-mMIMO maintains its advantage for the 95th-percentile users but at the cost of degraded median SE.

The use of high-gain directional communications is key to enable the capacity enhancement of millimeter-wave (mm-wave) for 5G-and-beyond networks. However, robust mm-wave coverage using directional beams entails significant beamsteering effort for maintaining precise beam alignment between the base station and the user. Numerous mm-wave beamsteering algorithms have been proposed in the literature, but verified largely using statistical channel models or limited measurements. Importantly this still leaves open the question of whether these studies can be directly translated to real outdoor mm-wave network deployments. In this paper, we present the results of the first extensive comparative evaluation of eight state-of-the-art mm-wave beamsteering algorithms based on outdoor mm-wave measurements using phased antenna arrays. We collect received signal strength data over fine-grained 3D angular orientations for 78 spatially-dense user positions in a European city, comprising an open-source dataset of over 421,000 individual measurements. We then perform an empirical evaluation of the beamsteering algorithms in the context of link establishment for static and link maintenance for mobile users. Overall, our results show that these state-of-the-art mm-wave beamsteering algorithms, when tested on real measurement data, perform far from optimal and worse than originally reported. A key takeaway from our evaluation is that the practical feasibility of beamsteering algorithms strongly depends on the trade-off between the incurred beam training delay and the antenna gains facilitated in the initial beam training stage which limit the established link budget. Our study of link maintenance for mobile users showed significant deviations from the maximum achievable performance for all algorithms, including those that leverage correlation of the mm-wave sparse link opportunities or historical link information for link recovery. This shows that, to enable seamless connectivity in future mm-wave networks, more sophisticated beamsteering algorithms must be designed that intelligently adapt to the site-specific mm-wave channel while taking into account the realistic antenna beams of commercially-viable phased antenna arrays.

To address the end-to-end implications of realistic beam management control operations, in this paper we present a full-stack 5G-NR-compliant ns-3 framework which implements realistic scheduling, transmission, and reception of 5G-NR downlink control signals, namely synchronization signal block (SSB) and channel state information-reference signal (CSI-RS). The framework is based on a non-standalone downlink 5G-NR millimeter wave (mm-wave) network, enabling control signal transmission via an LTE network connection for handover coordination and overcoming radio link failure (RLF). Our framework provides customizable interfaces to a threshold-based beam management operation. Furthermore, we propose three 5G-NR-compliant beam management strategies and analyze their performance using site-specific propagation and pedestrian mobility data. Overall, our results show the end-to-end implications of realistic beam management operations required to maintain high-rate mobile user performance, with frequent signal-to-noise ratio variations reflected as significant drops in the achievable throughput and high delay spikes. This highlights the importance of the presented ns-3 framework compared to the ideal, instantaneous beam scanning supported in existing ns-3 mm-wave modules, for realistic 5G-NR beam management evaluation. To this end, we provide open access to the ns-3 code to facilitate the research efforts of the wireless community.

The use of high-gain directional communications is key to enable the capacity enhancement of millimeter-wave (mm-wave) spectrum for 5G-and-beyond networks. However, this entails extensive beam management to continuously track the mobile user to maintain precise transmitter (TX)/receiver (RX) beam alignment. Detailed characterization of mm-wave link opportunities in real-world urban environments and corresponding beam management requirements, are key to addressing the great challenge of mobility support for mm-wave urban deployments. In this paper, we present the results of our large-scale outdoor urban mm-wave measurements using phased antenna arrays, where we collected received signal strength data over fine-grained 3D TX/RX orientations for 78 spatially-dense RX positions in three sub-regions in Aachen, totaling over 421,000 individual measurements. Our results show 2–8 spatial link opportunities per RX position, where only up to 40% of all RX orientations result in a feasible mm-wave connection. Our results suggest that the beam management burden would fall heavily on the mobile user-side, and be moderate on the TX-side, for a typical walk in vicinity of the serving cell. For example, the required TX-side beam steering from one RX position to the next to maintain a minimum data rate of 1 Gbps is less than 12° (twice the half power-beamwidth), but up to 88° at the RX-side. Therefore, obtaining spatially-dense and fine-grained angular propagation data is a critical input for the design and evaluation of mm-wave beam management protocols.

Increasingly more wireless devices adopt CSMA at the MAC layer, primarily due to the ubiquitous utilization of Wi-Fi, but also due to the development of new standards that implement CSMA variants, like LAA-LTE. It is thus critical to use reliable modelling tools that accurately evaluate the performance of large-scale CSMA deployments, as expected in practice. This paper compares the accuracy and computational efficiency of stochastic geometry modelling (SGM) and our proposed hybrid model, for estimating the MAC-layer throughput and signal-to-interference-and-noise ratio (SINR) of large-scale CSMA wireless deployments. Using ns-3 simulations as the baseline reference, we show that our hybrid model is significantly more accurate and at least one order of magnitude faster to compute than SGM. SGM underestimates the SINR, yet as it neglects the sensing overhead, it usually overestimates the MAC-layer throughput substantially. We thus argue that our hybrid model is preferable to SGM for modelling large-scale CSMA deployments.

Increasingly more unlicensed bands accommodate multiple broadband technologies, e.g. Wi-Fi and LTE in the 5 GHz band and Wi-Fi 6E and 5G NR-U in the new 6 GHz unlicensed band. In these bands, the technologies coexist primarily via distributed MAC spectrum sharing mechanisms, so it is crucial to evaluate their performance using the right tools that reliably and accurately model the coexisting schemes. CSMA and duty cycle are representative schemes in this context, but the few modelling tools that capture their coexistence performance are yet to be properly evaluated in terms of accuracy and computational efficiency. This paper is the first to compare three such modelling tools - stochastic geometry modelling (SGM), ns-3, and our hybrid model - which comprise different time granularities and abstraction levels for the MAC mechanisms. We compare both the accuracy of their estimated SINR and throughput versus ns-3 and their computational efficiency, for CSMA and duty cycle coexistence. Furthermore, we make recommendations pertinent to future broadband coexistence cases. Our results show that ns-3 captures important effects of the CSMA sensing time on the SINR, but its computation time may be prohibitive for modelling coexisting dense networks. This tool is thus useful primarily for studying dynamic MAC features in small to medium-sized coexisting deployments. SGM is the least accurate and also requires a long computation time, so we argue it is the least attractive approach to model coexisting CSMA and duty cycle technologies. Finally, our hybrid model yields the time-average per-link SINR with reasonable accuracy and the throughput with good accuracy and is at least two orders of magnitude faster to compute than SGM and ns-3, making it very well suited for extensive Monte Carlo simulations and large networks.

Exploiting multi-antenna technologies for robust beamsteering to overcome the effects of blockage and beam misalignment is the key to providing seamless multi-Gbps connectivity in millimeter-wave (mm-wave) networks. In this paper, we present the first large-scale outdoor mm-wave measurement study using a phased antenna array in a typical European town. We systematically collect fine-grained 3D angle-of-arrival (AoA) and angle-of-departure (AoD) data, totaling over 50,000 received signal strength measurements. We study the impact of phased antenna arrays in terms of number of link opportunities, achievable data rate and robustness under small-scale mobility, and compare this against reference horn antenna measurements. Our results show a limited number of 2-4 distinct spatial link opportunities per receiver location, indicating that the mm-wave multipath richness in a European town is surprisingly similar to that of dense urban metropolises. The results for the phased antenna array reveal that significant losses in estimated data rate occur for beam misalignments in the order of the half-power beamwidth, with significant and irregular variations for larger misalignments. By contrast, the loss for horn antennas is monotonically increasing with the misalignment. Our results strongly suggest that the effect of non-ideal phased antenna arrays must be explicitly considered in the design of agile beamsteering algorithms.

The promise of millimeter-wave (mm-wave) frequencies for high capacity cellular networks depends on precise alignment of the narrow directional beams to either line-of-sight (LOS) or strong non-LOS links. Given this sensitivity of mm-wave communication to the spatial distribution of LOS/NLOS links, site-specific propagation data from ray-tracing simulations is an important tool for mm-wave networking studies. In this paper, we present the first detailed validation of mm-wave ray-tracing against large-scale outdoor measurements. We consider fine-grained angle-of-arrival (AoA), angle-of-departure (AoD) and received signal strength (RSS) data, using both horn and phased array antennas. We show that ray-tracing captures well the distribution of multipath clusters (MPCs) in terms of number of MPCs per receiver location, AoA, AoD and individual MPC structure. Moreover, our results indicate that ray-tracing provides accurate propagation data based on publicly available 3D building models which lack detailed material properties. Overall, our results show that individual propagation paths can be accurately identified in the ray-tracing data, with a median RSS prediction error within 5 dB of the measured RSS for all MPCs. This is an encouraging result which confirms the viability of ray-tracing propagation data as an input for mm-wave networking studies, on e.g. beam management protocols.

The 3rd ACM Workshop on Millimeter-Wave Wireless Networks and Sensing Systems (mmNets'19) is focused on the design and implementation of new millimeter-wave (mmWave) protocols and systems that can enable multi-Gbps wireless connectivity for 5G-and-beyond cellular systems and wireless LANs, as well as new wireless sensing and imaging systems. The goal of the mmNets'19 workshop is to bring together researchers from mmWave hardware, communication and signal processing, wireless networking, and mobile applications to set the future research agenda of mmWave systems, and present innovative ideas that will help realize the vision of extremely high data rate wireless networks and novel advanced sensing applications. The workshop will serve as a platform for both academia and industry to identify key challenges, present solutions and advance the field of mmWave technology.

Recent technological advancements are making the use of compact, low-cost, low-power mm-wave radars viable for providing environmental awareness in a number of applications, ranging from automotive to indoor mapping and radio resource optimisation. These emerging use-cases pave the road towards networks in which a large number of radar and broadband communications devices coexist, sharing a common spectrum band in a possibly uncoordinated fashion. Although a clear understanding of how mutual interference influences radar and communications performance is key to proper system design, the core tradeoffs that arise in such scenarios are still largely unexplored. In this paper, we provide results that help bridge this gap, obtained by means of an analytical model and extensive simulations. To capture the fundamental interactions between the two systems, we study mm-wave networks where pulsed radars coexist with communications devices that access the channel following an ALOHA policy. We investigate the effect of key parameters on the performance of the coexisting systems, including the network density, fraction of radar and communication nodes in the network, antenna directivity, and packet length. We quantify the effect of mutual interference in the coexistence scenario on radar detection and communication network throughput, highlighting some non-trivial interplays and deriving useful design tradeoffs.

Due to the increasing density of wireless devices, the ever-growing demands for extremely high data rates, and the spectrum scarcity at the sub-6 GHz bands, making use of the spectrum-rich millimeter-wave (mmWave) frequencies is among the most important technology trends for future wireless networks. The major commercial potential of mmWave networks has led to mmWave being considered a key element for 5G-and-beyond mobile cellular networks, as well as for emerging Gbps-speed Wi-Fi networks based on the IEEE 802.11ad and draft IEEE 802.11ay standards. Despite this intense interest in mmWave communications from both the research community and industry, much fundamental research is still needed, especially at the higher layers of the networking stack.

Emerging Wi-Fi technologies are expected to cope with large amounts of traffic in dense networks. Consequently, proposals for the future IEEE 802.11ax Wi-Fi amendment include sensing threshold and transmit power adaptation, in order to improve spatial reuse. However, it is not yet understood to which extent such adaptive approaches - and which variant - would achieve a better balance between spatial reuse and the level of interference, in order to improve the network performance. Moreover, it is not clear how legacy Wi-Fi devices would be affected by new-generation Wi-Fi implementing these adaptive design parameters. In this paper we present a thorough comparative study in ns-3 for four major proposed adaptation algorithms and we compare their performance against legacy non-adaptive Wi-Fi. Additionally, we consider mixed populations where both legacy non-adaptive and new-generation adaptive populations coexist. We assume a dense indoor residential deployment and different numbers of available channels in the 5 GHz band, relevant for future IEEE 802.11ax. Our results show that for the dense scenarios considered, the algorithms do not significantly improve the overall network performance compared to the legacy baseline, as they increase the throughput of some nodes, while decreasing the throughput of others. For mixed populations in dense deployments, adaptation algorithms that improve the performance of new-generation nodes degrade the performance of legacy nodes and vice versa. This suggests that to support Wi-Fi evolution for dense deployments and consistently increase the throughput throughout the network, more sophisticated algorithms are needed, e.g. considering combinations of input parameters in current variants.

Object localization is at the core of several context-aware applications envisioned for the Internet of Things. However, the present localization approaches are often too expensive, or are limited by indoor layouts and noise. In recent years, radio tomographic imaging (RTI) has generated great interest as a device-free localization approach. While several RTI algorithms have been proposed in the literature, their robustness and comparative performance in indoor environments, with real-world impairments, has not yet been experimentally studied. In this paper, we compare the performance of three state- of-the-art RTI algorithms and analyze the impact of different environmental conditions and algorithm parameters on the localization accuracy. Our experimental results show that multipath propagation is the main limiting factor for indoor localization using RTI: our measurements over a diverse set of indoor locations exhibit a 90th percentile localization error of between 0.8 m and 2.85 m. Additionally, our experiments reveal that co-channel interference and external human mobility further degrade the accuracy by 5%-20%. Furthermore, we show that while some improvements are achieved through modifications in network configuration and the fundamental RTI algorithm, these changes (such as increased node density and multi-channel RTI) are not feasible for cost- effective deployments.

In its evolution to provide ever higher data rates, the Wi-Fi standard has incorporated sophisticated PHY-layer techniques, which has in turn increased the complexity of network-wide interference relationships. Proper modelling of the resulting inter-device interactions is crucial for accurately estimating Wi-Fi network performance, especially in the contemporary context of traffic and network densification. Event-driven simulators like the open-source ns-3 are in principle able to capture these interactions, however it is imperative to validate, against experimental results, whether their underlying models reflect the network behaviour in practice. In this paper we first perform experiments in a large-scale indoor testbed to validate the IEEE 802.11ac Wi-Fi model in ns-3, for various channel width and allocation configurations. Our results show that ns-3 captures Wi-Fi co-channel interactions with reasonable precision, but fails to model adjacent channel interference (ACI), which our experiments show to be critical in dense networks. We therefore propose and implement an ACI model in ns-3. Importantly, our model successfully captures the qualitative behaviour of the CSMA/CA mechanism when transmissions on adjacent channels occur. Further, our ACI implementation significantly improves the accuracy of both the network and per-device throughput estimates for the considered dense IEEE 802.11ac network compared to the basic ns-3 Wi-Fi model without ACI. For example, without ACI modelling, ns-3 overestimates the aggregate network throughput by up to 230%, whereas with our ACI implementation the aggregate throughput estimate is no more than 65% higher than the experimental results.

As a device-free approach, radio tomographic imaging (RTI) is ideally suited for low-cost indoor localization in context-aware Internet-of-Things applications. However, the fundamental RTI algorithm relies on shadowing of the line of sight (LOS) links and therefore, conventional RTI implementations using 2.4 GHz sensing networks (microRTI) fail to accurately localize users in multipath-rich indoor environments. The localization accuracy is further degraded by external human movement that affects the signal propagation. In this paper, we propose mmRTI, a novel RTI approach based on a highly-directional, millimeter-wave sensing network, that aims to improve indoor localization by utilizing the LOS-dominant nature of millimeter-wave signal propagation. We experimentally evaluate mmRTI, operating at 60 GHz, with and without human movement around the sensing network, in two indoor environments, and compare its performance against the conventional microRTI approach. We observe that mmRTI achieves a 90%-ile localization error of 0.07m-0.25m, an improvement of 2.41 m-2.60 m compared to microRTI, while remaining unaffected by external human movement, which degrades the microRTI localization accuracy by up to 1.2 m.

Cellular deployments in the spectrum-rich mm-wave bands are anticipated to provide multi-Gbps connectivity in future 5G networks. Although measurements have demonstrated the feasibility of outdoor mm-wave links, the network-level opportunities and challenges of mm-wave cellular deployments are yet to be fully understood. In this paper we present a study of mm-wave urban network coverage and robustness - i.e. handover and beamsteering opportunities for supporting user mobility and resolving blockage of directional mm-wave links by moving obstacles. We use real 3D building data with ray tracing and investigate the impact of base station placement and density, building materials, and antenna directivity. Our results show that mm-wave networks can provide multi-Gbps connectivity locally, but that over 25% of the network area remains in outage even with dense pico-cellular deployments and that supporting mobility is very challenging. Moreover, primary link blockage can be resolved via handover or beamsteering to a secondary link at only up to 40% and 25% of high-throughput locations, respectively. Given these networking challenges, we argue that multi-frequency heterogeneous networks are the key to exploiting the great promise of mm-wave for future 5G deployments, i.e. mm-wave as a hotspot capacity booster, rather than a solution for comprehensive cellular coverage.

IEEE 802.11ad millimeter-wave (mm-wave) WLAN nodes achieve high throughput at the cost of the frequent re-steering requirement of highly directional antenna beams to establish and maintain links disrupted by beam misalignment, node mobility, or link blockage. Consequently, rapid and robust beamsteering algorithms are essential to enable seamless communication in mm-wave WLANs. In this paper we propose AoASteer, a beamsteering algorithm that speeds up the link establishment process in IEEE 802.11ad mm-wave WLANs by preferentially searching over a subset of mm-wave antenna sectors predicted by 2.4 GHz angle of arrival (AoA) estimation at the access point (AP). We experimentally evaluate the performance of AoASteer through extensive measurements in several indoor and outdoor locations, using 60 GHz USRP-SiversIMA packet radio transceivers to gather real mm-wave link information, and a 2.4 GHz USRP-based receiver with an 8-element uniform linear antenna array for AoA estimation, both implemented using GNU Radio. Our evaluation results, obtained for APs with different numbers of beams per sector, show that AoASteer typically selects a near-optimal LOS or NLOS link to establish communication, while significantly reducing the link establishment latency. For example, for 4-sector mm-wave antennas, AoASteer reduces the latency by 46 μs for 73% of our measurement cases compared to the IEEE 802.11ad beamsteering algorithm, while achieving the highest data rate of 6.7 Gbps for 92% of the cases.

Wi-Fi is the dominant wireless indoor broadband solution and thus key for meeting the exponential traffic growth. The recent IEEE 802.11ac amendment enables PHY data rates exceeding 1 Gbps. However, it is not clear how this increased per-link performance, achieved especially via wider channels, translates to network-level performance. The latter is crucial for understanding the true potential of emerging Wi-Fi, as massive densification of network infrastructure is needed for keeping up with capacity demands. In this paper we present results from an extensive measurement study of the performance of IEEE 802.11ac Wi-Fi in a large 24-node indoor testbed. We investigate in detail the impact of channel width, network deployment density, and type/volume of traffic on the achieved network performance in dense indoor deployments. Our results show that using wide 80 MHz channels is beneficial only in dense networks with extremely high traffic loads, owing to strong adjacent channel interference (ACI) effects with narrower channels. We show that ACI not only reduces aggregate network throughput but causes severe unfairness among nodes, where up to half the nodes may experience starvation. Starvation occurs due to frequency flow-in-the-middle effects and the heterogeneous interference coupling among pairs of nodes typical of indoor deployments. However, our results also demonstrate that for bursty TCP traffic with loads typical of modern residential Wi-Fi deployments, there is no harm from using narrower 20 or 40 MHz channels. Finally, we study the scaling of IEEE 802.11ac deployments with increasing traffic load, which is highly relevant for the dimensioning of future networks. We show that ACI effects only become evident at very high loads which are beyond the per-node traffic demand expected in the foreseeable future. Therefore, in practice there is no network-level benefit from employing the wider channels enabled by IEEE 802.11ac, even in emerging highly dense indoor deployments.

The millimeter-wave (mm-wave) bands are currently being explored for multi-Gbps wireless local area networks (WLANs). Directional antennas are required to overcome the high attenuation inherent at the mm-wave frequencies. However, directionality makes link maintenance and establishment tasks complex, especially under node mobility, as slight misalignment of antenna beams between nodes leads to link disruption. Consequently, low latency beamsteering algorithms are needed for fast link re-establishment to support seamless data provisioning. Solutions based on exhaustive sequential scanning induce high latency, thereby disrupting communication. On the other hand, existing low latency proposals typically consider only static links, depend on additional hardware, or require a priori information about the network environment. In this paper, we propose a generic, fast mm-wave beamsteering algorithm that utilizes the previous valid link information to initiate the feasible antenna sector pair search and adaptively increases the sector search space around it to re-establish a link. Additionally, we experimentally evaluate the performance of our algorithm through measurements conducted in a real indoor environment using 60 GHz packet-radio transceivers. The results show that, compared to exhaustive sequential scanning, our algorithm reduces the required sector search space, and thereby the link re-establishment latency, by 89% on average compared to exhaustive sequential scanning.

Spectrum sharing mechanisms need to be carefully designed to enable inter-technology coexistence in the unlicensed bands, as these bands are an instance of a spectrum commons where highly heterogeneous technologies and deployments must coexist. Unlike in licensed bands, where multiple technologies could coexist only in a primary-secondary dynamic spectrum access mode, a spectrum commons offers competition opportunities between multiple dominant technologies, such as Wi-Fi and the recently proposed LTE in the 5 GHz unlicensed band. In this paper, we systematically study the performance of different spectrum sharing schemes for inter-technology coexistence in a spectrum commons. Our contributions are threefold. First, we propose a general framework for transparent comparative analysis of spectrum sharing mechanisms in time and frequency, by studying the effect of key constituent parameters. Second, we propose a novel throughput and interference model for inter-technology coexistence, integrating per-device specifics of different distributed MAC sharing mechanisms in a unified network-level perspective. Finally, we present a case study of IEEE 802.11n Wi-Fi and LTE in the 5 GHz unlicensed band, in order to obtain generalizable insight into coexistence in a spectrum commons. Our extensive Monte Carlo simulation results show that LTE/Wi-Fi coexistence in the 5 GHz band can be ensured simply through channel selection schemes, such that time-sharing MAC mechanisms are irrelevant. We also show that, in the general co-channel case, the coexistence performance of MAC sharing mechanisms strongly depends on the interference coupling in the network, predominantly determined by building shielding. We thus identify two regimes: (i) low interference coupling, e.g., residential indoor scenarios, where duty cycle mechanisms outperform sensing-based listen-before-talk (LBT) mechanisms and (ii) high interference coupling, e.g., open-plan indoor or outdoor hotspot scenarios, where LBT outperforms duty cycle mechanisms.

Recent rulings by US and UK regulators allowing access to unused portions of TV spectrum have elicited high hopes for the practical value of these TV whitespaces (TVWS) for secondary exploitation. However, this optimism has been largely fueled by rather simple early studies, which do not consider all the system aspects in sufficient detail; the few early experimental works reported are proprietary and consider only a small number of nodes, whereas the existing more sophisticated theoretical studies focus on singular aspects of secondary spectrum access in isolation from the overall system interactions. In this paper we study quantitatively the deployment of secondary networks in TVWS by considering two archetypal candidate systems: LTE-like cellular and Wi-Fi-like networks. We develop a systematic framework for the performance evaluation of secondary networks, which we then use to obtain realistic estimates of the performance of our example systems. The secondary network performance assessment methodology demonstrated in this paper can be directly applied for other regions and systems. Our work explicitly takes into account limitations arising from aggregate interference, user density, and specific secondary transceiver characteristics. We argue that a systematic analysis, jointly considering all of these aspects, is key for obtaining realistic and robust results on the estimated value of whitespace spectrum. Our detailed system-level approach reveals a much more conservative picture of the realistic benefit of TVWS deployments than what has been commonly assumed thus far. We find that cellular TVWS networks have limited capabilities, but that a macro-cellular-only network may be a viable option for traffic offloading. Our results also show that Wi-Fi-like secondary deployments in TVWS, although increasing coverage range, lead to increased congestion, which limits the system capacity.

Femtocell networks have emerged as an important means of offloading traffic from today's LTE cellular networks. However, deploying LTE femtocells increases the interference in the network, as they share licensed spectrum with the macro-cellular network. Consequently, a wide range of interference mitigation techniques have been proposed to enable coexistence of femtocells and macrocells. Techniques of various implementation and signaling complexity have been considered, but the majority of existing work evaluates their performance solely based on how well they mitigate interference for macrocell users. However, it is important that both macrocell and femtocell users achieve good performance. In this paper we present a comparative system-level study on the performance of several LTE femtocell frequency and power allocation schemes. We investigate whether more complex and dynamic resource allocation schemes, i.e. power control, are worthwhile from the point of view of femtocell networks compared to lower complexity static resource allocation schemes. We evaluate the performance in terms of the users' downlink throughput and achieved signal-to-interference-plus-noise ratio (SINR), and we consider different interference environments for a wide range of network densities (urban, suburban, and rural scenarios). Our results show that power control algorithms only yield a marginal increase in throughput for the LTE femtocell users. This suggests that increasing the complexity of static frequency allocation schemes by implementing power control algorithms is not justified.

An understanding of the underlying spatial structure of user-deployed wireless networks is invaluable for the design and optimization of opportunistic small-cell technologies in emerging multi-tier architectures. IEEE 802.11 Wi-Fi networks provide a real-world example of such a large-scale random network structure. In this paper we report on a high-resolution measurement survey of Wi-Fi deployments in Germany as a step towards deriving more sophisticated spatial models for transmitter distributions in emerging small-cell networks, which are the core of offloading strategies for future wireless Internet. Using a custom setup, our measurement campaign covered a range of urbanization and land use scenarios representative of Wi-Fi deployments in developed countries. Our data indicates a 14-fold increase in Wi-Fi density in urban residential areas over the last decade, consistent with increased broadband penetration. This supports our hypothesis that household density has become a strong predictor of access point (AP) density. We infer AP locations using a novel localization technique suitable for large measurement campaigns using off-the-shelf equipment. From this data, we derive pertinent spatial statistics to characterize the topologies of opportunistic networks. We derive nearest-neighbour distributions showing that realistic inter-neighbour distances are higher than predicted by a homogeneous Poisson Point Process, underlining the need for more refined spatial models than the random node location models which are currently prevalent. We also derive the two-pair correlation functions for the AP locations, showing that AP clustering is apparent over a large range of distances in our data set, which has implications for routing and interference mitigation.

The prospect of increasing wireless capacity via secondary access to spatio-temporally underutilized chunks of spectrum, so-called whitespaces, has been proposed as a central aspect of emerging radio systems in response to the imminent spectrum scarcity problem. In this demonstration we present a novel software tool that helps researchers, industry, and regulators in assessing the feasibility and value of secondary spectrum access beyond simple whitespace availability calculation. Whereas existing software applications merely provide visualization of estimated secondary spectrum over a geographic area, our tool uniquely enables a holistic evaluation of the realistic potential of whitespace technologies, by modelling the performance of entire secondary systems in the envisioned eco-system of dynamic spectrum access policy and technology. Our tool provides a unified and flexible software framework and assessment methodology to conduct such studies, and is composed of an extensive primary spectrum usage database, a graphical interface for user interaction, and an interface to an extensible MATLAB backend for numerical calculations. We showcase the deployment scenarios of cellular and Wi-Fi-like secondary networks in TVWS (TV whitespaces). We also compare the impact of employing FCC-type of regulatory rules (with a fixed power/no-talk distance configuration) against European WG-SE43 regulatory proposals (with probabilistic access and power control). The case studies we will demonstrate are based on real network configuration data of European and US TV networks.

The recent decisions by regulators in the USA and UK to open up unused portions of UHF spectrum for secondary use have been met with keen interest in using these TV white spaces (TVWS) for providing broadband services through Wi-Fi-like connectivity. Amid the ensuing media hype about “Wi-Fi on steroids”, there is a widespread perception that Wi-Fi operating in TVWS will provide much longer range, superior speeds, and more reliable connections than traditional Wi-Fi at 2.4 GHz. In this paper, we present a quantitative analysis of the performance of a network of Wi-Fi-like access points (APs) operating in TVWS in order to obtain a realistic estimate of the achievable range and downlink rate of such a secondary system. Unlike previous studies, we explicitly consider the effects of inter-AP interference and congestion and use real TVWS channel availability estimates from an example region of Germany. We confirm the favourable properties of the lower TVWS frequency range, of enabling better propagation through walls and a larger coverage range for the same power budget. Our results show that operating Wi-Fi hotspots in TVWS might be technologically attractive for outdoor rural areas where user demand is low. However, the extended coverage range in TVWS leads to increased congestion which rapidly limits the system capacity for an outdoor urban deployment with high user density. Therefore, a combined technological and economical analysis is essential before any final judgement can be reached about the viability of large-scale Wi-Fi deployments in TVWS.

Cooperative diversity can be applied to energy-constrained wireless sensor networks to significantly reduce node energy consumption. However, cooperation partners must be carefully selected and coordinated to practically exploit this energy saving potential. In this paper we investigate partner choice for energy efficient cooperation in a wireless sensor network. We formulate novel and computationally efficient partner choice heuristics for sensor nodes based on either global or local knowledge of average path loss values in the network. We present extensive simulation results of cooperation in a wireless sensor network to show that the proposed heuristics achieve near-optimally energy efficient partner selection. Our results also demonstrate that large network-wide energy savings are achieved as a result of cooperative communication. Therefore, our simple partner choice heuristics form the basis of an effective distributed cooperation protocol for improving the energy efficiency of a wireless sensor network. Very importantly from the point of view of practical implementation, we show that our partner choice heuristic based on local information is the most effective cooperation strategy for resource-constrained wireless sensor networks, as it yields superior energy conservation results while enabling fully distributed and scalable cooperation.

Energy efficient communication is a key requirement of energy-constrained wireless sensor networks. Cooperative diversity can be applied to wireless sensor networks to significantly reduce node energy consumption. However, judicious selection and coordination of cooperation partners is essential to exploiting this energy saving potential of cooperation. In this paper we investigate optimal partner choice and power allocation for energy efficient cooperation in wireless sensor networks. Our insights lead us to formulate power allocation and partner choice heuristics which form a simple and practical cooperation strategy for energy-constrained wireless sensor networks, allowing nodes to autonomously make near-optimal cooperation decisions. We show that the power allocation optimisation problem is non-linear, necessitating the use of a search to find the optimum solution. We present the resulting cooperative energy savings in terms of network geometry for a range of potential partner locations. Our results reveal that the partner-destination and the source-partner channels have roughly equal influence over cooperation decisions for optimal energy efficiency. We use this observation to devise a simple yet near-optimal power allocation heuristic.

This paper presents an investigation of the energy-efficiency of two major cooperative diversity techniques in short-range wireless sensor networks: virtual-MISO (multiple-input-single-output) and decode-and-forward. The total energy consumption of a cooperative system consists of the transmission energy, the transceiver circuit energy and the local communication energy cost. In virtual-MISO explicit local communication precedes the long-haul cooperative transmission, whereas in decode-and-forward nodes cooperate by overhearing and repeating each other's long-haul transmissions. Our energy analysis shows that decode-and-forward is overall the more energy-efficient cooperative scheme. We also show that cooperation can yield large energy savings compared to traditional SISO (single-input-single-output) communication, even when the local distance between cooperating nodes is increased and becomes comparable to the long-haul distance. Given specific local and long-haul distance ranges, we formulate the recommended cooperation strategy in terms of the number of cooperating partners to employ for optimal energy-efficiency.