Aerial wardens could pose significant security threats to satellite-ground communications due to their stronger received signals than legitimate ground users. To address this issue, the signals from all jamming satellites in low Earth orbit satellite networks, i.e., full jamming strategy (FJS), are utilized to counter the detection of the aerial warden. However, this worsens the communication quality of ground users. To improve it, we utilize the difference in visible spherical crowns between the ground user and the aerial warden due to the Earth blockage to propose the safeguard-zone strategy (SGS) via merely muting the jamming satellites visible to the ground user. To evaluate the effectiveness of the proposed strategies, we propose a stochastic geometry-based analytical framework to derive the covert probability and connection probability. To capture the trade-off between covertness and reliability, the effective covert rate, defined as the product of transmission rate, covert probability, and connection probability, is also analyzed and optimized. The results validate the accuracy of the analytical expressions and illustrate that SGS outperforms the FJS in the connection probability and effective covert rate with a small loss in covert probability, which can be compensated by increasing the transmit power or the number of jamming satellites.

As uncrewed aerial vehicles (UAVs) are widely used in many communication scenarios, the issue of security has also caused much concern due to the line-of-sight propagation. Therefore, in this paper, we study an air-to-ground covert communication system, where a UAV transmitter Alice transmits messages covertly to a ground receiver Bob under the communication behavior detection of a ground warden Willie with location uncertainty and concurrent interference from other ground environmental nodes whose locations follow a two-dimensional Poisson point process. Under this setup, we first provide the approximate probability distribution of the aggregated interference power from all environmental co-channel nodes to facilitate the covert analysis. Then, we derive the average covert probability separately for two cases: case 1 assumes that Willie knows the received power from Alice; case 2 assumes that Willie knows the probability distribution of the received power from Alice. Next, we derive the connection probability and the covert throughput which is the maximal transmission rate under the covertness and reliability constraints. Numerical results demonstrate the feasibility of air-to-ground covert communication with location and interference uncertainty. The results also show the average covert probability and covert throughput in case 2 are higher than that in case 1, especially for a sparse deployment of interfering nodes.

The combination of reconfigurable intelligent surface (RIS) and integrated sensing and communication (ISAC) can improve the resource utilization in non-line-of-sight scenarios. To sense the target with high accuracy, it is necessary to directionally reflect the sensing signal toward the sensing target via the RIS, improving the sensing performance. However, enhancing signal quality may increase the risk of information leakage when the sensing target is the warden. Against this background, we investigate a covert transmission problem in an RIS assisted ISAC system. Specifically, we obtain a tractable form of covertness constraint in terms of minimum detection error probability via the optimal detection threshold. Then, we maximize the sum covert transmission rate by jointly optimizing the beamforming of confidential signal and jamming signal as well as the RIS’s phase shift, while ensuring the reliability, covertness and sensing constraints. Owing to the effectiveness of the deep reinforcement learning algorithm in processing high-dimensional data and making intelligent decision, we propose a twins-deep deterministic policy gradient-based joint covert beamforming and the phase shift of RIS optimization (TD3-CBP) algorithm to solve the above non-convex problem. Finally, simulation results demonstrate the effectiveness of the proposed TD3-CBP algorithm in the covertness performance, achieving an average of 16.1 % higher sum covert transmission rate than the benchmark algorithms.

The long propagation delay makes the delay characteristics of data transmission in low Earth orbit (LEO) satellite networks limited by retransmission. Therefore, this paper focuses on the retransmission delay characteristics through analyzing the local delay, defined as the mean times required for the serving satellite successfully transmitting the message to the ground user. We propose a general analytical framework to evaluate the local delay in massive LEO satellite-to-ground downlink networks. Specifically, binomial point process is used to model the locations of satellites. Considering Nakagami fading and directional transmission, we derive the conditional success probability under a given network topology. On this basis, we first give an exact expression for the local delay and further provide an asymptotic analysis when the signal-to-interference-plus-noise ratio tends to zero and infinity. Additionally, we analyze the local delay in three special cases: noise-limited, Rayleigh fading and infinite antenna array of satellite. Numerical results verify our analysis and show that Rayleigh fading model and the asymptotic analysis can simplify and effectively approximate the exact result of the local delay under Nakagami fading.

LinkedIn, one of the world's largest platforms for professional networking and job seeking, encounters various modeling challenges in building recommendation systems for its job matching product, including cold-start, filter bubbles, and biases affecting candidate-job matching. To address these, we developed the STAR (Signal integration for Talent And Recruiters) system, leveraging the combined strengths of Large Language Models (LLMs) and Graph Neural Networks (GNNs). LLMs excel at understanding textual data, such as member profiles and job postings, while GNNs capture intricate relationships and mitigate cold-start issues through network effects. STAR integrates diverse signals by uniting LLM and GNN capabilities with industrial-scale paradigms including adaptive sampling and version management. It provides an end-to-end solution for developing and deploying embeddings in large-scale recommender systems. Our key contributions include a robust methodology for building embeddings in industrial applications, a scalable GNN-LLM integration for high-performing recommendations, and practical insights for real-world model deployment.

We present LinkSAGE, an innovative framework that integrates Graph Neural Networks (GNNs) into large-scale personalized job matching systems, designed to address the complex dynamics of LinkedIn's extensive professional network. Our approach capitalizes on a novel job marketplace graph, the largest and most intricate of its kind in industry, with billions of nodes and edges. This graph is not merely extensive but also richly detailed, encompassing member and job nodes along with key attributes, thus creating an expansive and interwoven network. A key innovation in LinkSAGE is its training and serving methodology, which effectively combines inductive graph learning on a heterogeneous, evolving graph with an encoder-decoder GNN model. This methodology decouples the training of the GNN model from that of existing Deep Neural Network (DNN) models, eliminating the need for frequent GNN retraining while maintaining up-to-date graph signals in near real-time, allowing for the effective integration of GNN insights through transfer learning. The subsequent nearline inference system serves the GNN encoder within a real-world setting, significantly reducing online latency and obviating the need for costly real-time GNN infrastructure. Validated across multiple online A/B tests in diverse product scenarios, LinkSAGE demonstrates marked improvements in member engagement, relevance matching, and member retention, confirming its generalizability and practical impact.

The combination of reconfigurable intelligent surface (RIS) and integrated sensing and communication (ISAC) can improve the resource utilization in non-line-of-sight scenarios. However, the private information in this situation raises security concerns when the transmission behavior is detected by wardens. Against this background, we investigate a covert transmission problem in an RIS assisted ISAC system. Specifically, we obtain a tractable form of covertness constraint in terms of minimum detection error probability via the optimal detection threshold, paving the way for optimization process. Then, the sum covert transmission rate is maximized by jointly optimizing the beamforming of confidential signal and jamming signal as well as the RIS’s phase shift. To solve the above non-convex problem, we propose a joint covert beamforming and the phase shift of RIS optimization-based twins-deep deterministic policy gradient (CBP-TD3) algorithm. Finally, simulation results demonstrate the effectiveness of the proposed CBP-TD3 algorithm in the covertness.

In this paper, we analyze the covert UAV-to-UAV (U2U) communication with the ground warden under a Poisson field of ground interferers and the blockage effect of air-to-ground propagations. With the aid of stochastic geometry, we derive the average covert probability and connection outage probability to quantify the covert communication performance for the scenarios with and without interference. By comparing the two scenarios, the introduction of interference increases the minimum average covert probability. Increasing both the density and transmission power of interferers can improve the average covert probability. However, the improvement of covertness is achieved by increasing the connection outage probability. To capture the competing requirements of covertness and reliability, we analyze the effective covert communication rate defined as the product of the average covert probability, connection success probability, and data transmission rate. The UAV transceiver can also flexibly raise its flight altitude to improve effective covert communication rate. Moreover, our results reveal that covert U2U communication performs better in high blockage environments such as dense urban. In summary, this study provides theoretical guidance for designing covert U2U communication systems.

In cellular networks, the performance of cell-edge users is notably deficient, especially for those receiving almost comparable signal strengths from the serving base station (BS) and interfering BS(s). To improve their performance, a cell-edge UAV deployment scheme is proposed, where the UAVs are deployed to hover over these cell-edge users to serve them. Specifically, the locations of BSs are modeled using a Poisson point process (PPP) and the Voronoi cells are formed. We consider two distinct types of cell-edge users positioned at the Voronoi vertices and boundaries, corresponding to the worst-case and boundary users, respectively. Due to the intractable spatial distribution of the interfering UAVs, we utilize the PPP and binomial point process (BPP) approximations to characterize their interference. Subsequently, we derive the success probabilities for these two types of cell-edge users. The results show the similar effectiveness of the two approximate models for the locations of the UAVs. Furthermore, we obtain the asymptotic success probabilities to analyze the performance in high-reliability regimes and propose an effective approximation based on the asymptotic behavior to simplify the analytical expressions. The results highlight the substantial performance enhancement through the proposed UAV deployment scheme for cell-edge users.

To address the poor performance experienced by cell-corner users located equidistantly to the serving base station (BS) and the nearest interfering BSs, this paper proposes a flexible and general air-ground cooperation scheme based on unmanned aerial vehicles and dynamic BS coordination in the form of BS silencing (BSS) or joint transmission (JT). To show the role of unmanned aerial vehicle (UAV) in the cooperation, we define the UAV-to-BS power ratio (UBPR) as a critical parameter to measure whether introducing the UAV can improve the user’s performance. Using stochastic geometry tools, we derive the success probability of the user located at the corner of the Voronoi diagram, called the worst-case user, in Poisson cellular networks. To facilitate the comparison between different modes of cooperation, we further analyze the cooperation gain including the diversity and power gains through the asymptotic outage probability in the high-reliability regime. Furthermore, to reflect the impact of the cooperation scheme on the overall network performance, we analyze the normalized spectral efficiency, which, unlike those adopted in existing works, accounts for the costs of both resource occupancy and data exchange. Numerical results validate the accuracy of our analytical findings and demonstrate the effectiveness of the proposed scheme for cell-corner users.

In this paper, we present LiGNN, a deployed large-scale Graph Neural Networks (GNNs) Framework. We share our insight on developing and deployment of GNNs at large scale at LinkedIn. We present a set of algorithmic improvements to the quality of GNN representation learning including temporal graph architectures with long term losses, effective cold start solutions via graph densification, ID embeddings and multi-hop neighbor sampling. We explain how we built and sped up by 7x our large-scale training on LinkedIn graphs with adaptive sampling of neighbors, grouping and slicing of training data batches, specialized shared-memory queue and local gradient optimization. We summarize our deployment lessons and learnings gathered from A/B test experiments. The techniques presented in this work have contributed to an approximate relative improvements of 1% of Job application hearing back rate, 2% Ads CTR lift, 0.5% of Feed engaged daily active users, 0.2% session lift and 0.1% weekly active user lift from people recommendation. We believe that this work can provide practical solutions and insights for engineers who are interested in applying Graph neural networks at large scale.

We present LiRank, a large-scale ranking framework at LinkedIn that brings to production state-of-the-art modeling architectures and optimization methods. We unveil several modeling improvements, including Residual DCN, which adds attention and residual connections to the famous DCNv2 architecture. We share insights into combining and tuning SOTA architectures to create a unified model, including Dense Gating, Transformers and Residual DCN. We also propose novel techniques for calibration and describe how we productionalized deep learning based explore/exploit methods.

In cellular networks, the performance of cell-boundary users with almost equal distance from the serving base station (BS) and the nearest interfering BS(s) is extremely poor. In order to improve the performance of these users, this paper proposes a cell-boundary UAV deployment scheme. To quantify the performance enhancement, the locations of BSs are modeled as a Poisson point process forming Voronoi cells. Two distinct types of cell-boundary users located at the vertices and boundaries of Voronoi cells are served by a UAV-mounted BS hovering over them, and the success probabilities of these two types of users are derived with the stochastic geometry tool. Furthermore, the asymptotic success probabilities are obtained to analyze the performance in high-reliability regime and an effective approximation is proposed based on the asymptotic behavior to simplify the performance evaluation. Finally, the re-sults highlight the substantial performance enhancement achieved through UAV-assisted communication for the cell-boundary users.

To address the poor performance experienced by cell-edge users located equidistantly to the serving base station (BS) and the nearest interfering BS(s), this paper proposes an air-ground coordinated multipoint scheme assisted by unmanned aerial vehicles and the dynamic coordinated BSs selected from the sets of the serving and the equidistant interfering BSs. Using stochastic geometry tools, we derive success probabilities in a Poisson cellular network for the users located at corners of the Voronoi diagram called worst-case users served using non-coherent joint transmission. To reflect the impact of the coordinated transmission on the overall network performance, we also deduce the normalized spectral efficiency. Numerical results validate the accuracy of our analytical findings and show the superior performance of the proposed scheme for the worst-case users.

Combining intelligent reflecting surface (IRS) with an unmanned aerial vehicle to form aerial IRS (AIRS) swarms is an effective way to enable panoramic full-angle reflection, high configuration flexibility and reliable air-ground line-of-sight (LoS) connections, especially in millimeter wave (mmWave) bands. In this paper, we use stochastic geometry to provide a performance analytical framework for AIRS swarm-assisted mmWave cellular networks, where each base station (BS) has an AIRS swarm to assist downlink communications. To capture the swarm property of AIRSs and the dependence between AIRS swarms and BSs, the AIRSs in each swarm are uniformly and randomly distributed in a finite circular area centered on their assisting BS. Incorporating different LoS/non-LoS propagation characteristics, a two-step user association policy is proposed to pursue an efficient communication link between the BS and its users with the assistance of AIRS swarm. We derive the coverage probability and area spectral efficiency (ASE) by considering three types of interference which are directly from interfering BSs, reflected by active AIRSs of other swarms and reflected by the assisting AIRS for the typical user, respectively. The results reveal that AIRS swarms enhance both coverage and ASE performance and have the optimal height and density that maximize the coverage probability in mmWave cellular networks.

To alleviate the time overhead of the handover and beam reselection caused by user mobility, a velocity-based association policy has been adopted in heterogeneous networks. However, it might introduce severe interference from cross-tier interfering base stations which are closer to the users than the serving one. To deal with this problem, we propose two cross-tier interference coordination (CTIC) schemes for mobile users in heterogeneous networks: one jointly considers the transmit power and path loss (TPL-CTIC), and the other further incorporates the directional antenna gain (TPG-CTIC). Considering the beam misalignment and the time overhead caused by user mobility, we derive the overall coverage probability and the average effective Shannon rate (ESR) of mobile users to characterize the user-perceived performance. Taking the cost of the proposed CTIC schemes into account, we further derive the average normalized throughput to evaluate the overall network performance. Numerical results demonstrate that compared with the case without CTIC, the proposed schemes significantly improve the overall coverage probability in the low signal-to-interference-plus-noise ratio regime and the average ESR. To capture the expense of proposed schemes, we define the average normalized throughput. Although TPL-CTIC performs better in terms of the overall coverage probability and the average ESR, TPG-CTIC outperforms in terms of the average normalized throughput.

Intelligent reflecting surface (IRS) has been regarded as an efficient technology to enhance the network performance. Since the IRS can assist transmitters to communicate in a directional reflection way, its deployment usually has a strong correlation with the locations of transmitters. In order to well capture the directional transmission and the spatial correlation while guaranteeing the analytical tractability, we propose an IRS-assisted network model based on a Gauss-Poisson process and a simple yet effective IRS reflected model. Under this setup, we derive the success probabilities for two association policies, namely the dedicated transmitter-receiver pairs with fixed-distance and the nearest associations. To highlight the impact of the IRS, we also derive the success probabilities of three special cases, i.e., without deploying IRSs, ignoring the IRS interference, and blocked direct link between transceivers. The results show that deploying IRS improves the success probability, especially in the cases of IRSs deployed near the transmitters and blocked direct link, and the IRS interference cannot be ignored which becomes the bottleneck factor in the case of transmitters equipped with massive antenna arrays. Owing to the proposed model, the impact of the correlation between IRSs and transmitters is well studied, which provides useful guidance for integrating IRSs into wireless networks.

Air platforms, such as unmanned aerial vehicles, airships, and balloons are expected to complement traditional ground networks to provide flexible coverage solutions. However, most existing models for air-ground integrated networks (AGINs) neglect the spatial dependence caused by the complementary deployment of the aerial and ground nodes. Accordingly, in this paper, we propose two AGIN models with horizontal dependence that differ in the vertical dimension, namely uniformly independent altitudes and location-dependent altitudes. The air platforms serve as aerial base stations, distributed as a marked Poisson hole process, and provide flexible beam coverage through varying altitudes. Under this setup, we propose a region-based user association scheme and derive the association probabilities as well as the serving distance distributions of an arbitrarily located user. Considering Nakagami fading and air-to-ground propagation properties, we characterize the signal-to-interference ratio and area spectral efficiency for each model. Using the proposed analytical framework, we demonstrate the importance of deploying the air platforms more sensibly to provide targeted services and flexible beam coverage in reducing the load of base stations and improving the user coverage and network capacity performance.

The low Earth orbit (LEO) satellite constellation in the terahertz (THz) band is one of the promising technologies to provide global coverage and high-rate transmission to the airplane, whose key performance evaluation is required before realistic deployment. In this paper, we consider a downlink THz LEO satellite-airplane network with directional antennas, where the satellites follow a binomial point process on a spherical surface. Incorporating the distinctive THz propagation properties and directional beams in the two scenarios with/without molecular absorption noise, we propose an analytical framework to evaluate the coverage probability and average achievable rate of an arbitrary airplane. The analytical results are expressed in terms of the Laplace transforms of the interference-plus-molecular absorption noise and interference. Specifically, considering a frequency reuse strategy, we first provide the probability distributions of the distances from an arbitrary satellite, the serving satellite and any non-serving satellite to the target airplane, and calculate the probabilities of whether an interfering satellite is blocked by the Earth. Numerical results verify the accuracy of the analytical expressions and show that fewer orthogonal channels and lower LEO satellite altitude lead to better coverage and rate performance. As a key parameter in network design, the number of satellites has two competing effects on coverage and rate, and an optimum satellite number exists. Overall, this study provides theoretical guidance for deploying and planning THz satellite constellations.

We introduce generalized deep mixed model (GDMix), a class of machine learning models for large-scale recommender systems that combines the power of deep neural networks and the efficiency of logistic regression. GDMix leverages state-of-the-art deep neural networks (DNNs) as the global models (fixed effects), and further improves the performance by adding entity-specific personalized models (random effects). For instance, the click response from a particular user m to a job posting j may consist of contributions from a DNN model common to all users and job postings, a model specific to the user m and a model specific to the job j. GDMix models not only possess powerful modeling capabilities but also enjoy high training efficiency especially for web-scale recommender systems. We demonstrate the capabilities by detailing their use in Feed and Ads recommendation at LinkedIn. The source code for the GDMix training framework is available at https://github.com/linkedin/gdmix https://github.com/linkedin/gdmix under the BSD-2-Clause License.

In millimeter-wave (mm-wave) cellular networks, directional antenna arrays are typically adopted to mitigate the severe propagation loss. However, the interference caused by such highly directional beams may, in turn, result in a significant number of transmission failures, especially for dense networks. To tackle this problem, we propose two inter-cell interference coordination (ICIC) schemes in mm-wave bands: one is merely based on the path loss incorporating the blockage effect (PL-ICIC); the other considers both path loss and directivity gain (PG-ICIC). To fully investigate both schemes, we first derive an exact expression for the success probability (reliability) of the typical (served) user. We further provide an asymptotic analysis for the success probability and propose an effective approximation based on the asymptotic signal-to-interference ratio (SIR) gain relative to no ICIC. Secondly, to incorporate the cost of ICIC schemes, we derive the approximate normalized throughput taking into account that some users cannot be served due to limited resources. Numerical results show that the two proposed schemes provide significant reliability improvements in the low-SIR regime, and the higher the number of antennas, the wider the SIR range for which there is an improvement. In addition, compared with PL-ICIC, PG-ICIC balances the available resources among all users well.

In UAV-assisted cellular networks (UACNs), the UAV tier should be flexibly adjusted to the terrestrial tier, which causes location dependence between UAVs and ground base stations (GBSs). To capture it, this paper proposes a 3D UACN model with inter-tier dependence and analyzes the performance of such UAV-assisted network via stochastic geometry. Specifically, UAVs with random altitudes are deployed outside the exclusion regions of GBSs to avoid the strong inter-tier interference, and accordingly, the locations of GBSs and UAVs follow a PPP and a marked Poisson hole process (MPHP), respectively. Under this setup, we propose a division region-based user association scheme, and provide analytical results for the association probability, link distance distribution, and success probability which are verified through simulations. Numerical results unveil that: 1) the exclusion region should be carefully designed to achieve the optimal access probability and 2) significant gains (especially for cell edge users) can be obtained if UAVs are introduced and deployed agilely.

This paper provides a deep analysis for millimeter-wave device-to-device (mm-wave D2D) underlaid cellular networks to investigate how the performance of individual users or link percentiles under constraints are affected by the interaction between D2D and cellular communication modes integrated with the salient mm-wave features. Under a general Nakagami fading model, we derive the upper bounds for the$b$-moments of the conditional signal-to-interference-noise ratio (SINR) distributions for the two modes given the network realization, and further calculate the upper bounds of SINR meta distributions (MDs). Moreover, we extend the MD-based analysis to evaluate the fine-grained performance for the transmission effectiveness, through deriving the bounding results for the user-perceived rate MD and spatial rate capacity (SRC), the latter of which is defined as the effective density of users satisfying a given transmission rate requirement with a certain probability. Numerical results validate the tightness of the proposed bounds and reveal the different effects of the network configuration parameter associated with mm-wave on the link and network performances from the view of individual users.

The information freshness, quantified by the emerging idea of Age of Information (AoI), has recently received widespread attentions from both academia and industry. However, most prior works on AoI considered the single source-destination pair model and provided the performance analysis merely through queueing-theoretic models, which do not lend themselves to the analysis of large-scale wireless networks. This paper focuses on the AoI analysis in a large-scale setting using random spatial models in stochastic geometry. Specifically, we first adopt rateless coding for packet delivery to improve the transmission reliability while reducing the transmission delay. Then, to quantify the enhancement of the information freshness by rateless coding, we derive the average peak age of information (PAoI), which characterizes the maximum value of AoI immediately before an update packet is received, accounting for the spatial variation of network nodes and temporal dynamics of service packets. Numerical results show the significant benefits of rateless codes relative to the commonly used fixed-rate codes in terms of the PAoI.

Despite of the promising benefits brought by device-to-device (D2D) technique, the attainable performance gains in an underlaying network may still be eroded by the severe mutual interference between cellular and D2D communications. To address this critical issue, we first propose an exclusion-based D2D activation (EDA) scheme with hybrid spectrum allocation to suppress the mutual interference and improve the user performance, where only the D2D users outside the exclusion regions are active and share the frequency band with the cellular users within the exclusion regions. Secondly, we establish a stochastic geometry-based analytical framework and analyze the success probabilities of both cellular and D2D users, where the distribution of the activated D2D pairs follows a bipolar Poisson hole process. The results demonstrate the significant performance improvement and the effectiveness on the interference suppression brought by the proposed EDA scheme and the hybrid spectrum allocation.

Although plenty of structural variant detecting approaches for human genomes can be looked up in the literatures, little has been acknowledged on the effectiveness of those structural variant softwares for plant genomes. Moreover, it has been demonstrated frequent occurrences for those structural variant detecting softwares to find too many false structural variants. In this paper, we devote to detect deletions, insertions, and inversions, in total of three kinds of structural variants occurring in Hedou12 genome in contrast to Williams82 genome. To find more potential structural variants, we try to develop new principles to detect discordant and split read map sets supporting structural variants. Aiming to enhance the precision of structural variant detections, we propose two new sequencing characteristic based probability models, which use the sequencing parameters of Hedou12 genome as well as the parameters for Hedou12 paired-end reads to be aligned onto Williams82, to evaluate the probability for a potential structural variant to occur in. To remove the false members from those potential structural variants, we propose a set cover problem model to describe formally on which potential structural variants it should accept to achieve as high as possible a probability summation. This will achieve a solution with more credible structural variants, which can be verified by comparing with DELLY version 0.5.8 and LUMPY version 0.2.2.3. Our algorithm has been verified to be able to find deletions, insertions, and inversions in Hedou12 in contrast to Williams82 DELLY as well as LUMPY fails to find.

In millimeter-wave (mm-wave) cellular networks, high-gain beamforming, realized with directional antenna arrays, is typically adopted to mitigate the severe propagation loss. However, the interference caused by such highly directional beams may, in turn, result in a significant number of transmission failures, especially for dense networks. To tackle this problem, we propose two inter-cell interference coordination (ICIC) schemes in mm-wave bands: one is merely based on the path loss incorporating the blockage effect (PL-ICIC); the other considers both path loss and directivity gain (PG-ICIC). To fully investigate the performance of both schemes, we first derive the exact expression for the success probability (reliability) of the typical user that is served. Secondly, to reflect the cost of interference coordination, we further derive the overall success probability taking into account that some users cannot be served due to limited resources. Numerical results show that both the proposed two schemes provide significant reliability improvements in the low signal-to-interference ratio (SIR) regime, in particular, the higher the number of antennas, the wider the range of SIR threshold for which there is an improvement. In addition, compared with PL-ICIC, PG-ICIC balances the available resources among all users well.

In 5G wireless communication network, massive machine type communication (mMTC) is an emerging research topic. For mMTC, non-orthogonal multiple access (NOMA) has been proposed to support its large-scale connectivity. Due to the sparsity of mMTC, compressed sensing based algorithms can be used to identify the active users and recover the sparse channel state information (CSI) vector. In this paper, we propose a Bayesian message passing algorithm based on expectation propagation (EP) for joint active user detection (AUD) and channel estimation (CE) in NOMA. The proposed method approximates the complicated target distribution with a Gaussian distribution to achieve linear complexity. Simulations demonstrate that the EP-based algorithm achieves better performance in joint AUD and CE than the exiting algorithms, especially in the low SNR regime.

Traditional low earth orbit mobile satellite systems (LEO-MSSs) cannot satisfy the demands of huge capacity in ultra-dense regions. In addition, due to the frequent passive group handover caused by high speed movement of LEO satellites, quality of service (QoS) can hardly be satisfied. In this paper, we propose an extensible multi-layer architecture model based on traditional LEO-MSSs to solve these problems. In the proposed architecture, we introduce high-altitude platforms (HAPs) and terrestrial relays (TRs) to cover ultra-dense regions and provide communication services. We also present the theoretical analysis of the system capacity and handover rate to evaluate the benefits brought by the proposed architecture. Moreover, we compare the system performance after adding HAPs and TRs in different scenarios. Simulation results validate that HAPs and TRs can increase capacity and reduce handover rate, and show how the capacity and handover rate are improved.

The actual antenna pattern model makes it difficult to investigate the performance of directional antenna arrays in the large-scale networks. In this paper, we propose a novel antenna pattern, named multi-level flat-top (MLFT) antenna pattern, to approximate the actual antenna pattern, and apply it to millimeter wave (mm-wave) cellular networks. To study the benefits of directional antenna arrays in depth, we derive the exact expression for the coverage probability of the typical user and its upper bound as well as their corresponding asymptotic expressions. The results show that the proposed antenna pattern provides a better tradeoff between the accuracy and the tractability in terms of both antenna pattern and theoretical analysis than previous models. In addition, the coverage probabilities for different antenna sizes can be asymptotically characterized by a horizontal shift (in dB) with each other for identical Nakagami fading parameter.

With the application of large-scale antenna systems in upcoming 5G networks, the overhead of channel measurement and feedback explodes exponentially. This paper aims to investigate an effective conversion method to converse the channel covariance from uplink to downlink for uniform planar array (UPA) and avoid the unnecessary overhead in FDD systems. First, a multiuser channel model is established according to the geometric structure of UPA. Then, by analyzing the relationship between the downlink channel covariance matrix (DCCM) from the uplink channel covariance matrix (UCCM), a general estimation algorithm is proposed for UPA. To reduce the computation complexity, a simple Kronecker algorithm is derived, which decomposes the DCCM of UPA to the DCCMs of the horizontal and vertical channel responses. Finally, both two algorithms are applied to the statistical beamforming algorithm in a multiuser FDD system with UPA. The Monte Carlo simulation verifies their validity.

Dense femtocell deployment in macrocell system is an important tendency in future cellular networks due to the increasing demand in indoor communications, where the cross-tier interference mitigation between the macrocell and femtocells is one of the most critical issues. However, the current studies mostly focus on the cross-tier interference in two-dimensional scenarios with a few works considering the tridimensional cell. In this paper, based on the traditional fractional frequency reuse (FFR) method, we propose a novel tridimensional frequency reuse (TFR) strategy to mitigate the cross-tier interference in two-tier femtocell networks. We analyze the performance of the proposed strategy and compare it with the traditional FFR method and the original co-channel method. Simulation results demonstrate the effectiveness of our proposed strategy on interference mitigation in the tridimensional cell.

The explosive growth of cellular network users and traffic stimulates the current network architecture to be more spectral-, energy-, and operation efficient and flexible. In this paper, we propose a transformium network architecture comprising transformium core, body and cloud, where the core dynamically schedules the body and cloud orchestrated with wireless network environment. To enable the transforming ability, some new emerging technologies, including software defined networking, network virtualization are employed in the proposed architecture. As a consequence, the transforimium network is a promising network architecture to improve the network capacity, quality of service and solve the network ossification problem.

Considering a heterogeneous cellular network where macro base stations (MBSs) are powered by grid power and small base stations (SBSs) are powered by renewable energy. We investigate the grid energy minimization problem by optimizing both the MBS active probability (MAP) and the SBS transmit power (STP) under the constraints of user quality of service (QoS) and renewable energy. Using tools from stochastic geometry, we first derive the expression of the service outage probability to evaluate user QoS. And then, the impacts of the MAP and the STP on the service outage probability are characterized. At last, the optimization problem minimizing the network grid energy consumption is formulated, and the closed expressions of the optimal MAP and STP are derived by investigating the monotonicity of the service outage probability. Numerical results show that our strategy of jointly optimizing the MAP and the STP is superior to the strategy of merely considering MBS sleeping with fixed SBS transmit power, and the superiority is more attractive with lower circuit power of SBS.

The current cellular networks have evolved to be more randomly, irregularly, and heterogeneously deployed to meet the exponential growth of mobile data traffic and the demand for seamless coverage, making the signal-to-interference ratio (SIR) distribution more challenging to analyze. Therefore, in this paper, we propose two simple approximative approaches to the SIR distribution of general heterogeneous cellular networks (HCNs) based on the ASAPPP method, which stands for “approximate SIR analysis based on the Poisson point process” and the MISR (mean interference-to-signal ratio)-based gain for each individual tier of the HCNs. Specifically, we first establish a per-tier ASAPPP approximation to general HCNs and then present an effective gain ASAPPP method as a further simplification when the path loss exponents are the same for all the tiers, that is, we give an explicit expression for the effective gain Geffof general HCNs such that the SIR distribution is obtained by scaling the SIR threshold θ to θ/Geff. The asymptotic behavior for the tail of the SIR distribution is also given. Furthermore, to highlight the simplicity and effectiveness of the approximative approaches, we derive the exact distribution of the SIR in the two-tier HCNs modeled by β-Ginibre and Poisson point processes and compare it with the approximate results. The results demonstrate that the proposed approaches give a simple yet excellent approximation for the SIR distribution.

The crushing demands for mobile data traffic drive the current cellular networks to become more heterogeneous, making the signal-to-interference ratio (SIR) distribution more difficult to analyze. In this paper we propose a simple approximative approach to the SIR distribution of heterogeneous cellular networks (HCNs) based on the ASAPPP method which stands for ``approximate SIR analysis based on the Poisson point process'' and the MISR (mean interference-to-signal ratio)-based gain for each individual tier of the HCN. The results demonstrate that this approach gives a tight approximation and asymptotically a lower bound for the coverage probability.