In the physical layer (PHY) of modern cellular systems, information is transmitted as a sequence of resource blocks (RBs) across various domains with each resource block limited to a certain time and frequency duration. In the PHY of 4G/5G systems, data is transmitted in the unit of transport block (TB) across a fixed number of physical RBs based on resource allocation decisions. Using sharp band-limiting in the frequency domain can provide good separation between different resource allocations without wasting resources in guard bands. However, using sharp filters comes at the cost of elongating the overall system impulse response which can accentuate inter-symbol interference (ISI). In a multi-user setup, such as in Machine Type Communication (MTC), different users are allocated resources across time and frequency, and operate at different power levels. If strict band-limiting separation is used, high power user signals can leak in time into low power user allocations. The ISI extent, i.e., the number of neighboring symbols that contribute to the interference, depends both on the channel delay spread and the spectral concentration properties of the signaling waveforms. We hypothesize that using a precoder that effectively transforms an OFDM waveform basis into a basis comprised of discrete prolate spheroidal sequences (DPSS) can minimize the ISI extent when strictly confined frequency allocations are used. Analytical expressions for upper bounds on ISI are derived. In addition, simulation results support our hypothesis.

This demo introduces WaveBox, a dynamic, software-defined waveform generation system developed to assess the resilience of communication networks against many types of interference scenarios. WaveBox features seamless waveform switching, allowing users to efficiently adjust interference patterns to adapt to diverse operational scenarios. We will showcase the system's effectiveness and versatility, highlighting its ability to adapt to evolving mission requirements. Additionally, the system's intuitive graphical user interface (GUI) supports rapid waveform adjustments, enhancing its responsiveness in dynamic environments. WaveBox can provide a flexible software-defined tool for evaluating the robustness of wireless systems.

Ensuring the safety and reliability of emerging Advanced Aerial Mobility (AAM) systems requires wireless communication to provide real-time monitoring and control information to ground stations. This greatly depends on the quality of the wireless links during aerial transit. In this demonstration, we present AirTwinX, designed to emulate flight control of flying vehicles while generating high-fidelity, context-aware models for air-to-air (AA) and air-to-ground (AG) communication links. Using GPU-accelerated ray tracing, AirTwinX can predict the quality of wireless links with near real-time updates based on the environmental geometry observed during flight. This data is then used to guide autonomous control decision-making. Additionally, the vehicle control toolchain employed in this work is based on software-in-the-loop (SITL) emulation of a commercial flight controller, enabling seamless translation of control policies from simulation to real-world hardware.

We design and implement a new multiple-input multiple-output (MIMO) frequency-division-duplex (FDD) spread-spectrum link and demonstrate real-time high-definition (HD) video streaming over a Radio Frequency System-on-a-Chip (RFSoC) software-radio testbed. To the best of our knowledge, this is the first-of-its-kind high-throughput low-latency full-duplex spread-spectrum link on RFSoC platforms which demonstrates an aggregated data throughput of 122 Mbps that supports real-time recording and playback of uncompressed full-HD video. The testbed comprises two Xilinx Zynq Ultrascale+ RFSoC ZCU111 evaluation kits with a custom-built application layer. A host-based graphical user interface (GUI) demonstrates live performance of the proposed 4×4 MIMO wireless link in terms of error vector magnitudepre-detection SINR and bit error rate (BER) and enables on-the-fly reconfiguration of link parameters such as spreading code sequence, transmit/receive antenna gains.

In highly dynamic and contested RF environments, communication systems must swiftly adapt to fluctuating spectral conditions while ensuring network quality of service (QoS). Maintaining link synchronization and spectral efficiency during waveform adaptation is particularly challenging due to the high mobility and autonomy of devices, coupled with the possibility of operating in GPS-denied environments. In this demo, we introduce a scalable, high-speed FPGA-based parallel decoding algorithm that leverages the HORNets signature adaptation protocol to address these challenges. Our solution preserves link synchronization within a multi-node network and enables efficient, lightweight waveform adaptation without reliance on GPS. The algorithm's resilience and effectiveness are demonstrated using a three-node cluster configuration, all subjected to non-colored or colored intentional interference.

Collaborative training methods like Federated Learning (FL) and Split Learning (SL) enable distributed machine learning without sharing raw data. However, FL assumes clients can train entire models, which is infeasible for large-scale models. In contrast, while SL alleviates the client memory constraint in FL by offloading most training to the server, it increases network latency due to its sequential nature. Other methods address the conundrum by using local loss functions for parallel client-side training to improve efficiency, but they lack server feedback and potentially suffer poor accuracy. We propose FSL-SAGE (Federated Split Learning via Smashed Activation Gradient Estimation), a new federated split learning algorithm that estimates server-side gradient feedback via auxiliary models. These auxiliary models periodically adapt to emulate server behavior on local datasets. We show that FSL-SAGE achieves a convergence rate of $\mathcal{O}(1/\sqrt{T})$, where $T$ is the number of communication rounds. This result matches FedAvg, while significantly reducing communication costs and client memory requirements. Our empirical results also verify that it outperforms existing state-of-the-art FSL methods, offering both communication efficiency and accuracy.

We consider the problem of interference avoidance in orthogonal frequency-division multiplexing (OFDM) communication systems. Unlike conventional OFDM systems that rely on static waveform designs, we propose to apply a coding sequence at the input of the inverse discrete Fourier transform (IDFT) operator to digitally shape the transmitted OFDM waveform. By dynamically optimizing the coding sequence to maximize the signal-to-interference-plus-noise ratio (SINR) at the receiver, the proposed approach enables resilient communication in heavily congested spectral environments in a technically simple and efficient manner. We carry out extensive simulation studies to evaluate the performance of the proposed system under various interference scenarios and demonstrate significant improvements in SINR when compared to conventional OFDM systems. When the proposed dynamic OFDM waveform optimization process is fielded in commercially available software-defined radio platforms, the transceiver can evade rapidly changing (msec–scale or faster) interference and survive in contested spectral environments.

Support vector machines (SVMs) have been the learning model of choice in numerous classification applications. While SVMs are widely successful in real-world deployments, they remain susceptible to mislabeled examples in training datasets where the presence of few faults can severely affect decision boundaries, thereby affecting the model's performance on unseen data. In this brief, we develop and describe in implementation detail a novel method based on $L_{1}$ -norm principal-component data analysis and geometry that aims to filter out atypical data instances on a class-by-class basis before the training phase of SVMs and thus provide the classifier with robust support-vector candidates for making classification boundaries. The proposed dataset curation method is entirely data-driven (touch-free), unsupervised, and computationally efficient. Extensive experimental studies on real datasets included in this brief illustrate the $L_{1}$ -norm curation method and demonstrate its efficacy in protecting SVM models from data faults during learning.

Physical layer key generation techniques aim to extract secret keys from the information contained in wireless channels. However, existing key generation schemes often rely on time-frequency domain waveforms for channel estimation, which not only makes secret extraction less reliable but may also compromise the confidentiality of the extracted secret information. This paper presents physical layer key generation methods relying on the Orthogonal Time Frequency and Space (OTFS) waveform. We present analysis showing that the delay-Doppler domain channel estimates obtained using OTFS are conducive to more secure and reliable secret extraction than time-frequency domain channel estimates obtained using the prevalent Orthogonal Frequency Division Multiplexing (OFDM). This analysis provides theoretical guarantees under certain simple assumptions. We then relax those assumptions in extensive time-division duplex (TDD) simulations and show that under realistic settings, OTFS offers the expected benefits to reliability and security. Our simulations show that the introduced OTFS schemes can reliably extract secret keys from channel estimates in scenarios where time-frequency domain methods deteriorate.

Directional antenna systems are gaining widespread adoption in wireless communication solutions, particularly using super-6 GHz bands in the electromagnetic spectrum. Hence, neighbor discovery and beam alignment in these directional wireless systems have attracted notable attention from researchers. However, fast neighbor discovery using directional wireless while maintaining covertness from eavesdroppers by minimizing the probability-of-intercept (POI) is an open research problem. We address this trade-off by proposing a sequential transceiver (or direction) selection protocol based on a tuning parameter (α) that guides the nodes to prioritize either goal by selecting the next operational transceiver subset for probing. We consider a 2-D multi-sector directional wireless system with electronic steering of transmission among sectors, assuming each sector is equipped with a transceiver and the sectors collectively cover the 2-D 360°horizon around the node. We design a time-slotted neighbor discovery protocol that employs a probabilistic approach to select only one transceiver to use for the next time interval. By changing α, we study its capability to control the probing direction and impact on neighbor discovery speed and POI. Results show random selection offers fastest neighbor discovery but increases vulnerability to passive eavesdropping. Choosing from transceivers placed opposite to the active one accelerates discovery, while prioritizing the adjacent ones enhances covertness. During the search for the best α, we observe that exclusively prioritizing either rapid discovery or minimizing POI for a long stretch does not yield an optimal solution.

In this work, we present a new federation framework for Union-Labs, an innovative cloud-based resource-sharing infrastructure designed for next-generation (NextG) and Internet of Things (IoT) over-the-air (OTA) experiments. The framework aims to reduce the federation complexity for testbeds developers by automating tedious backend operations, thereby providing scalable federation and remote access to various wireless testbeds. We first describe the key components of the new federation framework, including the Systems Manager Integration Engine (SMIE), the Automated Script Generator (ASG), and the Database Context Manager (DCM). We then prototype and deploy the new Federation Plane on the Amazon Web Services (AWS) public cloud, demonstrating its effectiveness by federating two wireless testbeds: i) UB NeXT, a 5G-and-beyond (5G+) testbed at the University at Buffalo, and ii) UT IoT, an IoT testbed at the University of Utah1.

We consider the problem of dynamically optimizing a multiple-input multiple-output (MIMO) wireless waveform in a given potentially heavily utilized fixed frequency band with applications in near-field or far-field autonomous machine-to-machine communications. In particular, we find the transmitter beam weight vector and the pulse code sequence that maximize the signal-to-interference-plus-noise ratio (SINR) at the output of the maximum SINR joint space-time receiver filter. We propose and derive two novel model-based solutions: (a) Disjoint, space first (transmit weight vector) then time (pulse code sequence) waveform optimization and (b) jointly optimal transmit weight vector and pulse code sequence optimization (a mixed integer programming problem.) The proposed formally derived algorithmic solutions are studied in extensive simulations under varying waveform code length, near-field/far-field and spread-spectrum/ non-spread-spectrum interference, in light and dense interference scenarios. Our findings highlight the effectiveness of the described methods compared to static conventionally designed MIMO links and the remarkable ability of the joint space-time optimized waveforms to avoid heavy interference.

By applying a simple shortest/minimum-cost routing algorithm, the mobile ad-hoc network (MANET) with heavy data transmissions may be easily congested if multiple routes meet at the same relay node. Therefore, those busy nodes should be avoided when a new path is established. The task of optimal path seeking becomes more challenging when a MANET is equipped with directional antennas that may cause directional interference with neighboring receivers. The motivation of our research is to build an intelligent proactive routing scheme for MANETs with directional antennas. Our directional routing protocol considers not only the global traffic distribution in different areas of the MANET, but also the properties of directional antennas. It uses a spatio-temporal deep learning algorithm to predict the next-time snapshot of a directional heat map (DHM), which shows the traffic density distribution in each network location as well as the coverage of each directional antenna. The DHM is then used to identify the optimal path that can avoid congested areas as well as the interference from all neighboring directional links. Furthermore, an optimization algorithm is designed to perform optimal path selection. It splits a single path into multiple paths converge later on into one path, if the path needs to go around a congested area. Therefore, our routing scheme achieves better quality-of-service (QoS) performance than existing routing schemes.

Decentralized bilevel optimization has received increasing attention recently due to its foundational role in many emerging multi-agent learning paradigms (e.g., multi-agent meta-learning and multi-agent reinforcement learning) over peer-to-peer edge networks. However, to work with the limited computation and communication capabilities of edge networks, a major challenge in developing decentralized bilevel optimization techniques is to lower sample and communication complexities. This motivates us to develop a new decentralized bilevel optimization called DIAMOND (decentralized single-timescale stochastic approximation with momentum and gradient-tracking). The contributions of this paper are as follows: i) our DIAMOND algorithm adopts a single-loop structure rather than following the natural double-loop structure of bilevel optimization, which offers low computation and implementation complexity; ii) compared to existing approaches, the DIAMOND algorithm does not require any full gradient evaluations, which further reduces both sample and computational complexities; iii) through a careful integration of momentum information and gradient tracking techniques, we show that the DIAMOND algorithm enjoys $\mathcal{O}\left( {{ \in ^{ - 3/2}}} \right)$ in sample and communication complexities for achieving an ϵ-stationary solution, both of which are independent of the dataset sizes and significantly outperform existing works. Extensive experiments also verify our theoretical findings.

Key predistribution schemes are promising lightweight solutions to be placed as the cornerstone of key management systems in multihop wireless networks. The intermediate decryption–encryption problem, however, is considered as the security threat of such schemes. Multipath algorithms have been proposed to face such a shortcoming. Alas, these solutions are vulnerable against the node capture attack, where the attacker compromises a fraction of network nodes. In this article, we propose BlocKP, a Blockchain-based solution to increase the resistance of the network against the node capture attack. BlocKP utilizes disjoint key paths for a key-exchange process, where the keying materials form a block at the source side. Each key path step generates the next block of the Blockchain until the keying materials reach the destination. BlocKP is a general framework applicable to any key predistribution schemes. We propose BlocKP in two versions BlocKP-I and BlocKP-II, where the latter enhances the resistance of BlocKP-I using erasure codes at the cost of negligible control traffic. We analytically show that BlocKP improves the resistance of the network against the node capture attack to almost perfect resistance, using just a small number of paths. We evaluate our solution by performing extensive simulations, considering three baseline key predistribution schemes, including probabilistic asymmetric key predistribution (PAKP), strong Steiner trade (SST), and unital key predistribution (UKP). We equipped these schemes with a compatible multipath algorithm to offer end-to-end security. Results show that BlocKP improves the throughput up to 5% and decreases the flow completion time into 20% compared to baseline schemes. It has comparable routing traffic, latency, and throughput with augmented solutions but up to 60% improvement in the resistance against the node capture attack.

Reconfigurable intelligent surface (RIS) offers tremendous spectrum-and-energy efficiency in wireless networks. With the agility and mobility of an unmanned aerial vehicle (UAV), RIS can be mounted on a UAV to enable three-dimensional (3D) signal reflections, reliable air-ground connections, and higher configuration flexibility. However, the scalability of the aperture gain and the spatial multiplexing could not be guaranteed in a single UAV-enabled aerial RIS due to UAV’s limited payload and line-of-sight-dominated air-ground connection. In this paper, we study a UAV swarm-enabled aerial RIS (SARIS)-assisted downlink communication system. The objective of the considered SARIS system is to maximize the weighted sum-rate of ground users by designing the transmit beamforming at the base station (BS), the phase shifts of SARIS reflecting elements, and SARIS 3D placement. For joint BS and SARIS beamforming design, we introduce two beamforming schemes with low computational complexity. For SARIS placement design, the optimal SARIS 3D position is obtained by leveraging the tools from stochastic geometry and considering the distributions of ground users. Simulation results confirm the validity of the analytical derivations. In particular, the SARIS placement plays a vital role in the system performance when the distances between users and the BS increase.

To mitigate the long-term spectrum crunch problem, the FCC recently opened up the 6 GHz frequency band for unlicensed use. However, the existing spectrum sharing strategies cannot support the operation of access points in moving vehicles such as cars and UAVs. This is primarily because of the directionality-based spectrum sharing among the incumbent systems in this band and the high mobility of the moving vehicles, which together make it challenging to control the cross-system interference. In this paper, we propose SwarmShare, a mobility-resilient spectrum sharing framework for swarm UAV networking in the 6 GHz band. We first present a mathematical formulation of the SwarmShare problem, where the objective is to maximize the spectral efficiency of the UAV network by jointly controlling the flight and transmission power of the UAVs and their association with the ground users, under the interference constraints of the incumbent system. We find that there are no closed-form mathematical models that can be used to characterize the statistical behaviors of the aggregate interference from the UAVs to the incumbent system. Then we propose a data-driven three-phase spectrum sharing approach, including Initial Power Enforcement, Offline-dataset Guided Online Power Adaptation, and Reinforcement Learning-based UAV Optimization. We validate the effectiveness of SwarmShare through an extensive simulation campaign. Results indicate that, based on SwarmShare, the aggregate interference from the UAVs to the incumbent system can be effectively kept below the target level without requiring the real-time cross-system channel state information. The mobility resilience of SwarmShare is also validated in coexisting networks with no precise UAV location information.

Federated learning (FL) is an emerging machine learning paradigm, in which clients jointly learn a model with the help of a cloud server. A fundamental challenge of FL is that the clients are often heterogeneous, e.g., they have different computing powers, and thus the clients may send model updates to the server with substantially different delays. Asynchronous FL aims to address this challenge by enabling the server to update the model once any client’s model update reaches it without waiting for other clients’ model updates. However, like synchronous FL, asynchronous FL is also vulnerable to poisoning attacks, in which malicious clients manipulate the model via poisoning their local data and/or model updates sent to the server. Byzantine-robust FL aims to defend against poisoning attacks. In particular, Byzantine-robust FL can learn an accurate model even if some clients are malicious and have Byzantine behaviors. However, most existing studies on Byzantine-robust FL focused on synchronous FL, leaving asynchronous FL largely unexplored. In this work, we bridge this gap by proposing AFLGuard, a Byzantine-robust asynchronous FL method. We show that, both theoretically and empirically, AFLGuard is robust against various existing and adaptive poisoning attacks (both untargeted and targeted). Moreover, AFLGuard outperforms existing Byzantine-robust asynchronous FL methods.

In this article we explore new techniques that can enable open remote experimentation for mobile networks. We first propose a cloud-based framework called CloudRAFT, based on which experimenters are allowed to remotely access and control experimental resources via public cloud AWS and share the resulting data and code via the cloud. Then, we discuss the enabling techniques for CloudRAFT, including Amazon serverless service, VNC-based remote command line, and Websocket-based real time communications, among others. Finally, we showcase the application of these techniques in enabling remote access to UB NeXT, a software-defined testbed that has been developed at University at Buffalo for wireless mobile network modeling, optimization and deployment. This work verifies the feasibility of accessing, controlling and sharing wireless testbeds through a remote public cloud.

Data-driven control based on AI/ML techniques has a great potential to enable zero-touch automated modeling, optimization and control of complex wireless systems. However, it is challenging to collect network traces in the real world because of high time and labor cost, weather limitations as well as safety concerns. In this work we attempt to tackle this challenge by designing a multi-fidelity simulator taking wireless Unmanned Aerial Vehicle (UAV) networks into consideration. We design the simulator by interfacing two Unmanned Aerial System (UAS) simulators we have developed in prior years: UBSim and UB-ANC. The former focuses on UAV network optimization and policy training by considering explicitly the network environments such as blockage dynamics, while the latter focuses more on high-fidelity UAV flight control. We first develop a coordination interface referred to as SimSocket for signaling exchanges between UBSim and UB-ANC in simulations, and then showcase coordinated simulations based on UBSim and UB-ANC. The new research that can be enabled by the integrated simulator is also discussed for digital twin-based UAS systems.

A node equipped with a multi-beam antenna can achieve a throughput of up to$m$times as compared to a single-beam antenna, by simultaneously communicating on its$m$non-interfering beams. However, the existing multi-beam medium access control (MAC) schemes can achieve concurrent data communication only when the transmitter nodes are locally synchronized. Asynchronous packet arrival at a multi-beam receiver node would increase the node deafness and MAC-layer capture problems, and thereby limit the data throughput. This paper presents an asynchronous multi-beam MAC protocol for multi-hop wireless networks, which makes the following enhancements to the existing multi-beam MAC schemes (i) A windowing mechanism to achieve concurrent communication when the packet arrival is asynchronous, (ii) A smart packet processing mechanism which reduces the node deafness, hidden terminals and MAC-layer capture problems, and (iii) A channel access mechanism which decreases resource wastage and node starvation. Our proposed protocol also works in the networks that deploy the nodes equipped with single-beam as well as multi-beam antennas. Simulation results demonstrate a superior performance of our proposed protocol.

We introduce RF-SITL, a radio frequency (RF) software-in-the-loop (SITL) channel emulator developed with GNU Radio and the University at Buffalo’s Airborne Networking and Communications (UB-ANC) emulator to enable integrated simulation of systems comprising multiple unmanned aerial vehicles (UAVs) interacting over a wireless communication channel. RF-SITL could be paired with any multi-robot simulator to enable I/Q sample-level fidelity simulation of communication interactions between the robots by accurately simulating channel effects, including interference, noise, distance-dependent path loss, and packet losses. RF-SITL works as follows: 1) it instantiates a virtual software-defined transceiver in GNU Radio for each UAV simulated in the UB-ANC Emulator; 2) it builds an interference channel model in which each network node receives the superposition of signals transmitted from other nodes; and 3) it synchronizes the location of each simulated UAV in the UB-ANC Emulator with the virtualized RF transceivers in RF-SITL, such that the communication channel between nodes can accurately model distance-dependent channel effects, such as path loss. With these capabilities, we can use both off-the-shelf and custom-built signal processing flowgraphs that simulate Gaussian Minimum Shift Keying (GMSK), 802.11-like Orthogonal Frequency Division Multiplexing (OFDM), and direct sequence spread-spectrum (DSSS) links in GNU Radio to simulate swarm UAV networks prior to their deployment in software-defined radios in a swarm UAV network.

This paper aims at designing high-data-rate swarm UAV networks with distributed beamforming capabilities. The primary challenge is that the beamforming gain in swarm UAV networks is highly affected by the UAVs’ flight altitude, their movements and the resulting intermittent link blockages, as well as the availability of channel state information (CSI) at individual UAVs. To address this challenge, we propose FlyBeam, a learning- based framework for joint flight and beamforming control in swarm UAV networks. We first present a mathematical formulation of the control problem with the objective of maximizing the throughput of swarm UAV networks by jointly controlling the flight and distributed beamforming of UAVs. Then, a distributed solution algorithm is designed based on a combination of Echo State Network learning and online reinforcement learning. The former is adopted to approximate the utility function for individual UAVs based on online measurements, by jointly considering the unknown blockage dynamics and other factors that affect the beamforming gain. The latter is used to guide the exploitation and exploration in FlyBeam. The effectiveness of FlyBeam is evaluated through an extensive simulation campaign. Results indicate that significant (up to 450%) beamforming gain can be achieved by FlyBeam. We also investigate the effects of blockages and UAV flight altitude on the beamforming gain. It is found that, which is somewhat surprising, higher (rather than lower) beamforming gain can be achieved by FlyBeam with denser blockages in swarm UAV networks.

Existing routing protocols may not be applicable in UAV networks because of their dynamic network topology and lack of accurate position information. In this paper, an enhanced flooding-based routing protocol is designed based on random network coding (RNC) and clustering for swarm UAV networks, enabling the efficient routing process without any routing path discovery or network topology information. RNC can naturally accelerate the routing process, with which in some hops fewer generations need to be transmitted. To address the issue of numerous hops and further expedite routing process, a clustering method is leveraged, where UAV networks are partitioned into multiple clusters and generations are only flooded from representatives of each cluster rather than flooded from each UAV. By this way, the amount of hops can be significantly reduced. The technical details of the introduced routing protocol are designed. Moreover, to capture the dynamic network topology, the Poisson cluster process is employed to model UAV networks. Afterwards, stochastic geometry tools are utilized to derive the distance distribution between two random selected UAVs and analytically evaluate performance. Extensive simulation studies are conducted to prove the validation of performance analysis, demonstrate the effectiveness of our designed routing protocol, and reveal its design insight.

Network-consensus-based decentralized learning optimization algorithms have attracted a significant amount of attention in recent years due to their rapidly growing applications. However, most of the existing decentralized learning algorithms could not achieve low sample and communication complexities simultaneously - two important metrics in evaluating the trade-off between computation and communication costs of decentralized learning. To overcome these limitations, in this paper, we propose a triple hybrid decentralized stochastic gradient descent (TH-DSGD) algorithm for efficiently solving non-convex network-consensus optimization problems for decentralized learning. We show that to reach an ϵ2-stationary solution, the total sample complexity of TH-DSGD is O(ϵ-3) and the communication complexity is O(ϵ-3), both of which are independent of dataset sizes and significantly improve the sample and communication complexities of the existing works. We conduct extensive experiments with a variety of learning models to verify our theoretical findings. We also show that our TH-DSGD algorithm is stable as the network topology gets sparse and enjoys better convergence in the large-system regime.

Decentralized nonconvex optimization has received increasing attention in recent years in machine learning due to its advantages in system robustness, data privacy, and implementation simplicity. However, three fundamental challenges in designing decentralized optimization algorithms are how to reduce their sample, communication, and memory complexities. In this paper, we propose a gradient-tracking-based stochastic recursive momentum (GT-STORM) algorithm for efficiently solving nonconvex optimization problems. We show that to reach an ϵ2-stationary solution, the total number of sample evaluations of our algorithm is Õ(m1/2ϵ-3) and the number of communication rounds is Õ(m1/2ϵ-3), which improve the O(ϵ-4) costs of sample evaluations and communications for the existing decentralized stochastic gradient algorithms. We conduct extensive experiments with a variety of learning models, including non-convex logistical regression and convolutional neural networks, to verify our theoretical findings. Collectively, our results contribute to the state of the art of theories and algorithms for decentralized network optimization.

To mitigate the long-term spectrum crunch problem, the FCC recently opened up the 6 GHz frequency band for unlicensed use. However, the existing spectrum sharing strategies cannot support the operation of access points in moving vehicles such as cars and UAVs. This is primarily because of the directionality-based spectrum sharing among the incumbent systems in this band and the high mobility of the moving vehicles, which together make it challenging to control the cross-system interference. In this paper we propose SwarmShare, a mobility-resilient spectrum sharing framework for swarm UAV networking in the 6 GHz band. We first present a mathematical formulation of the SwarmShare problem, where the objective is to maximize the spectral efficiency of the UAV network by jointly controlling the flight and transmission power of the UAVs and their association with the ground users, under the interference constraints of the incumbent system. We find that there are no closed-form mathematical models that can be used characterize the statistical behaviors of the aggregate interference from the UAVs to the incumbent system. Then we propose a data-driven three-phase spectrum sharing approach, including Initial Power Enforcement, Offline-dataset Guided Online Power Adaptation, and Reinforcement Learning-based UAV Optimization. We validate the effectiveness of SwarmShare through an extensive simulation campaign. Results indicate that, based on SwarmShare, the aggregate interference from the UAVs to the incumbent system can be effectively controlled below the target level without requiring the real-time cross-system channel state information. The mobility resilience of SwarmShare is also validated in coexisting networks with no precise UAV location information.

Federated learning (FL) is a prevailing distributed learning paradigm, where a large number of workers jointly learn a model without sharing their training data. However, high communication costs could arise in FL due to large-scale (deep) learning models and bandwidth-constrained connections. In this paper, we introduce a communication-efficient algorithmic framework called CFedAvg for FL with non-i.i.d. datasets, which works with general (biased or unbiased) SNR-constrained compressors. We analyze the convergence rate of CFedAvg for non-convex functions with constant and decaying learning rates. The CFedAvg algorithm can achieve an $\mathcal{O}\left( {1/\sqrt {mKT} + 1/T} \right)$ convergence rate with a constant learning rate, implying a linear speedup for convergence as the number of workers increases, where K is the number of local steps, T is the number of total communication rounds, and m is the total worker number. This matches the convergence rate of distributed/federated learning without compression, thus achieving high communication efficiency while not sacrificing learning accuracy in FL. Furthermore, we extend CFedAvg to cases with heterogeneous local steps, which allows different workers to perform a different number of local steps to better adapt to their own circumstances. The interesting observation in general is that the noise/variance introduced by compressors does not affect the overall convergence rate order for non-i.i.d. FL. We verify the effectiveness of our CFedAvg algorithm on three datasets with two gradient compression schemes of different compression ratios.

With wearable devices such as smartwatches on the rise in the consumer electronics market, securing these wearables is vital. However, the current security mechanisms only focus on validating the user not the device itself. Indeed, wearables can be (1) unauthorized wearable devices with correct credentials accessing valuable systems and networks, (2) passive insiders or outsider wearable devices, or (3) information-leaking wearables devices. Fingerprinting via machine learning can provide necessary cyber threat intelligence to address all these cyber attacks. In this work, we introduce a wearable fingerprinting technique focusing on Bluetooth classic protocol, which is a common protocol used by the wearables and other IoT devices. Specifically, we propose a non-intrusive wearable device identification framework which utilizes 20 different Machine Learning (ML) algorithms in the training phase of the classification process and selects the best performing algorithm for the testing phase. Furthermore, we evaluate the performance of proposed wearable fingerprinting technique on real wearable devices, including various off-the-shelf smartwatches. Our evaluation demonstrates the feasibility of the proposed technique to provide reliable cyber threat intelligence. Specifically, our detailed accuracy results show on average 98.5 percent, 98.3 percent precision and recall for identifying wearables using the Bluetooth classic protocol.

In addition to traditional networking devices (e.g., gateways, firewalls), current corporate and industrial networks integrate resource-limited Internet of Things (IoT) devices like smart outlets and smart sensors. In these settings, cyber attackers can bypass traditional security solutions and spoof legitimate IoT devices to gain illegal access to the systems. Thus, IoT device-class identification is crucial to protect critical networks from unauthorized access. In this paper, we propose Z-IoT, the first fingerprinting framework used to identify IoT device classes that utilize ZigBee and Z-Wave protocols. Z-IoT monitors idle network traffic among IoT devices to implement signature-based device-class fingerprinting mechanisms. Utilizing passive packet capturing techniques and optimal selection of filtering criteria and machine learning algorithms, Z-IoT identifies different types of IoT devices while guaranteeing the anonymity of the network data. To test Z-IoT's efficacy, we implemented several testbeds, including a total of 39 commodity IoT devices that communicate over ZigBee and Z-Wave protocols. Our experimental results showed an excellent performance in identifying different classes of IoT devices with average precision and recall of over 91%. Finally, the proposed framework yields no overhead to the IoT devices or the network traffic.

Network-distributed optimization has attracted sig-nificant attention in recent years due to its ever-increasing applications. However, the classic decentralized gradient descent (DGD) algorithm is communication-inefficient for large-scale and high-dimensional network-distributed optimization problems. To address this challenge, many compressed DGD-based algorithms have been proposed. However, most of the existing works have high complexity and assume compressors with bounded noise power. To overcome these limitations, in this paper, we propose a new differential-coded compressed DGD (DC-DGD) algorithm. The key features of DC-DGD include: i) DC-DGD works with general SNR-constrained compressors, relaxing the bounded noise power assumption; ii) The differential-coded design entails the same convergence rate as the original DGD algorithm; and iii) DC-DGD has the same low-complexity structure as the original DGD due to a self-noise-reduction effect. Moreover, the above features inspire us to develop a hybrid compression scheme that offers a systematic mechanism to minimize the communication cost. Finally, we conduct extensive experiments to verify the efficacy of the proposed DC-DGD and hybrid compressor.

The emergence of multi-beam directional antennas (MBDAs) has paved the way for fast and high-throughput data communications by providing concurrent multi-directional transmissions. However, the existing routing protocols are not capable of utilizing the advantages of MBDAs. In this paper, we have developed a new routing scheme, called volcano routing, which can exploit the concurrent packet dispatching capability of MBDAs for high-throughput data delivery. Its topology resembles the flow of volcano lava and several routing “pipes” are used, which can detour around the network “holes” or blocked areas. The routing process consists of two phases: 1) Main path search phase: There is a main path at the core of each pipe. Multiple optimal main paths are formed that have a high potential of adding side nodes to enable multi-beam communications. A hierarchical scoring system and the performance metrics are used to evaluate the quality of the main paths. 2) Volcano establishment phase: The top-quality main paths are selected, and side paths are formed around each main path to establish the volcano pipes. A multi-beam traffic scheduling and dispatching policy is also proposed to achieve better performance. Our results show that the volcano routing scheme can exploit the advantages of MBDAs for achieving high data rates.

Spatial spectrum sensing (SSS) enables mobile devices to sense the spatial spectrum holes and reuse the scarce spectrum opportunistically. In this paper, we model and analyze the SSS in uplink two-tier user-centric deployed heterogeneous networks (HetNets) where secondary users (SUs) sense the spectrum holes of cellular users. In the two-tier user-centric deployed HetNets, small cell base stations (SBSs) are deployed in hotspots with high user density, and macro base stations (MBSs) are deployed uniformly. Based on the semi-static power control mechanism, the average transmit power of cellular users associated with MBS and SBS are derived, respectively. Furthermore, the spatial false alarm probability and the spatial miss detection probability of a typical SU are obtained, respectively. Moreover, we characterize the coverage probability and the area spectral efficiency (ASE) of SU and cellular networks. The SUs' optimal SSS radius is obtained to maximize the ASE of the entire network while guaranteeing the ASE of cellular networks above a certain threshold. Simulation results show that when the density of SUs is small, a decrease in SUs' SSS radius reduces the coverage probability of SUs. However, it improves the ASE of SUs networks, although the inter-SU interference increases.

Unmanned aerial vehicles (UAVs) are becoming important communication infrastructures. One major challenge of communications with UAV networks is the routing protocol design. Due to the inherent characteristics (e.g., dynamic network topology and limited UAV device capabilities), it is difficult to directly apply existing routing protocols that utilize network topology information and routing path explorations. In this article, two novel routing protocols are designed based on random network coding (RNC) for a swarm UAV network, where UAVs operate cooperatively as a swarm, enabling efficient routing process. The first routing protocol utilizes the unique feature of RNC: Original packets can be decoded as long as an UAV accumulates sufficient generations. This property can be used to effectively expedite the underlying routing process. The second routing protocol further improves the efficiency where each forwarding UAV only needs to create a new generation rather than decoding original packets. Accordingly, the duration of each hop can be significantly reduced. Extensive simulations have been conducted to evaluate the performance of the designed routing protocols. The simulation results demonstrate that our designed routing protocols can effectively enhance the performance on both average transmission delay and delay violation probabilities compared to benchmark methods.

This paper develops a novel framework for the modeling and analysis of spatial spectrum sensing (SSS) for device-to-device (D2D) communications in uplink two- tier user-centric deployed heterogeneous networks (HetNets), where small cell base stations (SBSs) are deployed in the places with high user density termed hotspots introduced by 3GPP. We study the average transmit power of uplink users, the probability of spatial false alarm and the probability of spatial miss detection of a typical D2D transmitter (D2D-Tx) during SSS. Based on the results, we further characterize the coverage probability of a typical D2D user and the area spectral efficiency (ASE) of D2D networks. Simulation results verify our analysis and demonstrate the advantages of SSS-based D2D communications in future wireless networks.

Routing protocol design is one of the major challenges for swarm UAV networks. Due to the characteristics of a dynamic network topology, the low-complexity and the large volume of UAV devices, existing routing protocols based on network topology information, and routing table updates are not applicable in swarm UAV networks. In this paper, a Random Network Coding (RNC) enabled routing protocol is proposed to support an efficient routing process, which does not require network topology information or pre-determined routing tables. With the proposed routing protocol, the routing process could be significantly expedited, since each forwarding UAV may have already overheard some encoded packets in previous hops. As a result, some hops may be required to deliver a few encoded packets, and less hops may need to be completed in the whole routing process. The corresponding simulation study is conducted, demonstrating that our proposed routing protocol is able to facilitate a more efficient routing process.

Network consensus optimization has received increasing attention in recent years and has found important applications in many scientific and engineering fields. To solve network consensus optimization problems, one of the most well-known approaches is the distributed gradient descent method (DGD). However, in networks with slow communication rates, DGD's performance is unsatisfactory for solving high-dimensional network consensus problems due to the communication bottleneck. This motivates us to design a communication-efficient DGD-type algorithm based on compressed information exchanges. Our contributions in this paper are three-fold: i) We develop a communication-efficient algorithm called amplified-differential compression DGD (ADC-DGD) and show that it converges under any unbiased compression operator; ii) We rigorously prove the convergence performances of ADC-DGD and show that they match with those of DGD without compression; iii) We reveal an interesting phase transition phenomenon in the convergence speed of ADC-DGD. Collectively, our findings advance the state-of-the-art of network consensus optimization theory.

Massive MIMO and millimeter-wave communication (mmWave) have recently emerged as two key technologies for building 5G wireless networks and beyond. To reconcile the conflict between the large antenna arrays and the limited amount of radio-frequency (RF) chains in mmWave systems, the so-called hybrid beamforming becomes a promising solution and has received a great deal of attention in recent years. However, existing research on hybrid beamforming focused mostly on the physical layer or signal processing aspects. So far, there is a lack of theoretical understanding of how hybrid beamforming could affect mmWave network optimization. In this paper, we consider the impacts of hybrid beamforming on utility-optimality and queuing delay in mmWave cellular network optimization. Our contributions in this paper are three-fold: i) we develop a joint hybrid beamforming and congestion control algorithmic framework for mmWave network utility maximization; ii) we reveal a pseudoconvexity structure in the hybrid beamforming scheduling problem, which leads to simplified analog beamforming protocol design; and iii) we theoretically characterize the scalings of utility-optimality and delay with respect to channel state information (CSI) accuracy in digital beamforming.

In this paper, a delay-minimal joint framing and scheduling policy is proposed for cognitive sensor networks, where a secondary sensor node collects measurement samples, combines them together into packets and transmits the packets to a central data fusion center after proper scheduling, when a shared channel is released by primary nodes. The main objective of this study is to minimize the end-to-end delivery time for secondary sensor nodes based on the current input traffic rate, channel availability process and channel bit error probability. The proposed method outperforms conventional constant-length framing policies for any choice of packet length by minimizing the delay and preventing potential queue instability under dynamic channel conditions. This method can be utilized by secondary nodes in a wide variety of wireless sensing applications in order to collect time-sensitive data with minimal delays.

Massive MIMO and millimeter-wave communication (mmWave) have recently emerged as two key technologies for building 5G wireless networks and beyond. To reconcile the conflict between the large antenna arrays and the limited amount of radio-frequency (RF) chains in mmWave systems, the so-called hybrid beamforming becomes a promising solution and has received a great deal of attention in recent years. However, existing research on hybrid beamforming focused mostly on the physical layer or signal processing aspects. So far, there is a lack of theoretical understanding on how hybrid beamforming could affect mmWave network optimization. In this paper, we consider the impacts of hybrid beamforming on utility-optimality and queueing delay in mmWave cellular network optimization. Our contributions in this paper are three-fold: i) we develop a joint hybrid beamforming and congestion control algorithmic framework for mmWave network utility maximization; ii) we reveal a pseudoconvexity structure in the hybrid beamforming scheduling problem, which leads to simplified analog beamforming protocol design; and iii) we theoretically characterize the scalings of utility-optimality and delay with respect to channel state information (CSI) accuracy in digital beamforming.

To support the multi-gigabit per second data rates of 5G wireless networks, there have been significant efforts on the research and development of massive MIMO (M-MIMO) technologies at the physical layer. So far, however, the understanding of how M-MIMO could affect the performance of network control, and optimization algorithms remain rather limited. In this paper, we focus on analyzing the performance of the queue-length-based joint congestion control and scheduling framework over M-MIMO cellular networks with limited channel state information (CSI). Our contributions in this paper are twofold. First, we characterize the scaling performance of the queue-lengths and show that there exists a phase transitioning phenomenon in the steady-state queue-length deviation with respect to the CSI quality (reflected in the number of bits B that represent CSI). Next, we characterize the congestion control rate scaling performance and show that there also exists a phase transitioning phenomenon in steady-state congestion control rate deviation with respect to the CSI quality. Collectively, the findings in this paper advance our understanding of the tradeoffs between delay, throughput, and the accuracy/complexity of CSI acquisition in M-MIMO cellular network systems.

Conventional wireless mesh network (WMN)routing protocols are designed for the nodes that use the omni-directional or single-beam directional antennas. This research presents a throughput-efficient routing scheme for WMN, by taking advantage of the nodes equipped with the multi-beam directional antennas (MBDAs). Our routing design has the following two novel features: First, it is a cross-layer design by integrating the routing scheme with multi-beam oriented medium access control (MAC) scheme. Second, the routing topology has a diamond-like shape and uses the multi-path routes (i.e., one main path and a few side paths). The diamond shape makes the traffic converge and diverge periodically in the routing paths, which exploits the simultaneous data delivery capability of multi-beam antennas, and enhances the network throughput. Our simulation results demonstrate the high throughput efficiency of the proposed multi-beam routing scheme.

The last decade has seen significant advances in optimization-based resource allocation and control approaches for wireless networks. However, the existing work suffer from poor performance in one or more of the metrics of optimality, delay, and convergence speed. To overcome these limitations, in this paper, we introduce a largely overlooked but highly effective heavy-ball optimization method. Based on this heavy-ball technique, we develop a cross-layer optimization framework that offers utility-optimality, fast-convergence, and significant delay reduction. Our contributions are three-fold: i) we propose a heavy-ball joint congestion control and routing/scheduling framework for both single-hop and multi-hop wireless networks; ii) we show that the proposed heavy-ball method offers an elegant three-way trade-off in utility, delay, and convergence, which is achieved under a near index-type simple policy; and more importantly, iii) our work opens the door to an unexplored network control and optimization paradigm that leverages advanced optimization techniques based on “memory/momentum” information.

The Wireless Network after Next (WNaN) radio is a multi-transceiver multi-frequency mobile ad-hoc network system that features four independent frequency agile transceivers and uses a variation of Carrier Sense Multiple Access/Collision Avoidance (CSMA/CA). Utilizing WNaN's multiple transceivers, dynamic frequency assignment algorithm, and varying the hello packet interval can help alleviate the increased number of collisions due to the presence of an advantaged node such as an Unmanned Aerial Vehicle (UAV). Validation experiments in the field were completed to show that WNaN's features can be exploited to increase the packet delivery ratio and decrease the average packet delay when an advantaged node is present.

In recent years, there have been significant efforts on the research and development of Massive MIMO (M-MIMO) technologies at the physical layer. So far, however, the understanding of how M-MIMO could affect the performance of network control and optimization algorithms remains rather limited. In this paper, we focus on analyzing the performance of the queue-length-based joint congestion control and scheduling framework (QCS) over M-MIMO cellular networks with limited channel state information (CSI). Our contributions in this paper are two-fold: i) We characterize the scaling performance of the queue-lengths and show that there exists a phase transitioning phenomenon in the steady-state queue-length deviation respect to the CSI quality (reflected in the number of bits B that represent CSI); and ii) We characterize the congestion control rate scaling performance and show that there also exists a phase transitioning phenomenon in steady-state congestion control rate deviation respect to the CSI quality. Collectively, the findings in this paper advance our understanding of the trade-offs between delay, throughput, and the accuracy/complexity of CSI acquisition in M-MIMO cellular network systems.

In this paper we consider the problem of resource management for a Direct Sequence Code Division Multiple Access (DS-CDMA) wireless Visual Sensor Network (VSN) in a generalized fading environment. In a VSN application, the primary goal is ensuring that maximum video quality is achieved in spite of the prevailing network resource constraints. The Nash Bargaining Solution (NBS) was used in determining the transmission power and source and channel coding rates for each node. The nodes in the network negotiate in order to determine their transmission parameters. The task is to optimize the transmission powers (which are continuous) and the source and channel coding rates (which are discrete) for all the network nodes. Particle Swarm Optimization (PSO) is used to solve the mixed-integer optimization that arises. The analysis was carried out for a myriad of wireless multipath fading environments using a unified moment generating function (MGF) approach.

The problem of resource management in a Direct Sequence Code Division Multiple Access (DS-CDMA) wireless Visual Sensor Network (VSN) with M-array Phase Shift Keying (MPSK) modulation in an Additive White Gaussian Network (AWGN) channel was considered in this paper. Achieving maximum video quality, in spite of the prevailing network resource constraints, is of utmost importance in VSN applications. Our optimization scheme is based on the Nash Bargaining Solution (NBS). The nodes in the network negotiate in order to determine their transmission parameters (transmission powers; source and channel coding rates for each node). The task is to optimize the transmission powers (which are continuous) and the source and channel coding rates (which are discrete) for all the network nodes, while taking advantage of the improved bandwidth spectral efficiency provided by the higher order constellation.

In the present paper, we deal with the problem of allocating the network resources in multi-rate Direct Sequence Code Division Multiple Access (DS-CDMA) Visual Sensor Networks (VSNs). We consider a single-cell system where each node uses the same chip rate, but can transmit at a different bit rate. In wireless VSNs, we face the constraints of limited power lifetime and of an error-prone environment, mainly due to attenuation and interference. The proposed cross-layer scheme enables the Centralized Control Unit (CCU) to jointly allocate the transmission power, the transmission bit rate and the source-channel coding rates for each VSN node in order to optimize the delivered video quality. The transmission power of each visual sensor assumes values from a continuous range, while the rest of the resources take values chosen from an available discrete set. The numerical results demonstrate the performance of the proposed multi-rate scheme vs a single-rate system.

In this paper, we propose an approach to manage network resources for a direct sequence code division multiple access (DS-CDMA) visual sensor network where nodes monitor scenes with varying levels of motion. It uses cross-layer optimization across the physical layer, the link layer, and the application layer. Our technique simultaneously assigns a source coding rate, a channel coding rate, and a power level to all nodes in the network based on one of two criteria that maximize the quality of video of the entire network as a whole, subject to a constraint on the total chip rate. One criterion results in the minimal average end-to-end distortion amongst all nodes, while the other criterion minimizes the maximum distortion of the network. Our experimental results demonstrate the effectiveness of the cross-layer optimization.

We propose a game-theory-based cross-layer optimization scheme for wireless Direct Sequence Code Division Multiple Access (DS-CDMA) visual sensor networks. The scheme uses the Nash Bargaining Solution (NBS), which assumes that the nodes negotiate, with the help of a centralized control unit, on how to allocate resources. The NBS takes into account the video quality each node could achieve without making an agreement. The cross-layer optimization scheme determines the source coding rate, channel coding rate, and transmission power for each node. We compare the proposed game-theory-based scheme with competing schemes that minimize the average or maximum distortion among the nodes. Experimental results are presented and conclusions are drawn.