Federated learning (FL) has provided a new methodology for coordinating a group of clients to train a machine learning model collaboratively, bringing an efficient paradigm in edge intelligence. Despite its promise, FL faces critical challenges in Internet of Things (IoT) networks, particularly the combined impact of data heterogeneity and communication delays. This paper examines the theoretical and empirical impact of these factors in Asynchronous Federated Learning (AFL). Initially, we categorize existing memoryless asynchronous strategies as Asynchronous Updates with Delayed Gradients (AUDG). Our theoretical analysis of AUDG reveals the coupling effect of delays exacerbate the adverse impact of data heterogeneity, causing the global model to drift towards frequently active clients. To address this, we propose a gradient reusing mechanism, termed Pseudo-Synchronous Updates by Reusing Delayed Gradients (PSURDG). By leveraging storage to reuse historical gradients, PSURDG effectively decouples the correlation between delay and data heterogeneity. Crucially, we conducted a comprehensive convergence analysis covering both convex and non-convex settings, confirming the algorithm’s effectiveness in diverse optimization landscapes. Finally, both schemes are validated through rigorous analysis and extensive simulations on multiple datasets. The results demonstrate a clear trade-off that AUDG remains efficient under low data heterogeneity, while PSURDG improves convergence in high-heterogeneity scenarios with moderate delays, thereby providing a theoretical guideline for aggregation strategy selection in different edge intelligence scenarios.

Cellular-connected UAV systems have enabled a wide range of low-altitude aerial services. However, these systems still face many challenges, such as frequent handovers and the inefficiency of traditional transport protocols. To better study these issues, we develop a modular and scalable simulation platform specifically designed for UAVs communication leveraging the research ecology in wireless communication of MATLAB. The platform supports flexible 5G NR node deployment, customizable UAVs mobility models, and multi-network-interface extensions. It also supports multiple transport protocols including TCP, UDP, QUIC, etc., allowing to investigate how different transport protocols affect UAVs communication performance. In addition, the platform includes a handover management module, enabling the evaluation of both traditional and learning-based handover strategies. Our platform can serve as a testbed for the development and evaluation of advanced transmission strategies in cellular-connected UAV systems.

Event cameras offer unique advantages in scenarios involving high speed, low light, and high dynamic range, yet their asynchronous and sparse nature poses significant challenges to efficient spatiotemporal representation learning. Specifically, despite notable progress in the field, effectively modeling the full spatiotemporal context, selectively attending to salient dynamic regions, and robustly adapting to the variable density and dynamic nature of event data remain key challenges. Motivated by these challenges, this paper proposes EventMG, a lightweight, efficient, multilevel Mamba-Graph architecture designed for learning high-quality spatiotemporal event representations. EventMG employs a multilevel approach, jointly modeling information at the micro (single event) and macro (event cluster) levels to comprehensively capture the multi-scale characteristics of event data. At the micro-level, it focuses on spatiotemporal details, employing State Space Model (SSM) based Mamba, to precisely capture long-range dependencies among numerous event nodes. Concurrently, at the macro-level, Component Graphs are introduced to efficiently encode the local semantics and global topology of dense event regions. Furthermore, to better accommodate the dynamic and sparse characteristics of data, we propose the Spatiotemporal-aware Event Scanning Technology (SEST), integrating the Adaptive Perturbation Network (APN) and Multidirectional Scanning Module (MSM), which substantially enhances the model's ability to perceive and focus on key spatiotemporal patterns. By employing this novel collaborative paradigm, EventMG demonstrates the ability to effectively capture multi-level spatiotemporal characteristics of event data while maintaining a low parameter count and linear computational complexity, suggesting a promising direction for event representation learning.

Long-range depth is crucial for roadside perception, which helps vehicles detect potential threats earlier, respond promptly, and avoid collisions. However, a notable challenge with existing roadside perception methods is their difficulty in accurately perceiving objects at long-range depth. To this end, we propose a Priori Geometric-Guided long-range Depth estimation framework, named PGDepth. First, inspired by the human ability to perceive depth by referencing objects, we utilize priori Geometric as reference information for road objects, assigning each pixel a predefined depth range. Second, a coarse-to-fine approach is introduced to continuously refine the accuracy of depth distribution for pixels. Furthermore, we propose a new loss function to effectively supervise the depth distributions of road objects. Extensive experimental results on the DAIR dataset demonstrate that the proposed method surpasses previous state-of-the-art competitors.

Event cameras provide exceptionally high temporal resolution in dynamic vision systems due to their unique event-driven mechanism. However, the sparse and asynchronous nature of event data makes frame-based visual processing methods inappropriate. This study proposes a novel framework, Event-based Graph Spatiotemporal Sensitive Transformer (EGSST), for the exploitation of spatial and temporal properties of event data. Firstly, a well-designed graph structure is employed to model event data, which not only preserves the original temporal data but also captures spatial details. Furthermore, inspired by the phenomenon that human eyes pay more attention to objects that produce significant dynamic changes, we design a Spatiotemporal Sensitivity Module (SSM) and an adaptive Temporal Activation Controller (TAC). Through these two modules, our framework can mimic the response of the human eyes in dynamic environments by selectively activating the temporal attention mechanism based on the relative dynamics of event data, thereby effectively conserving computational resources. In addition, the integration of a lightweight, multi-scale Linear Vision Transformer (LViT) markedly enhances processing efficiency. Our research proposes a fully event-driven approach, effectively exploiting the temporal precision of event data and optimising the allocation of computational resources by intelligently distinguishing the dynamics within the event data. The framework provides a lightweight, fast, accurate, and fully event-based solution for object detection tasks in complex dynamic environments, demonstrating significant practicality and potential for application.

Federated learning has provided a new approach to coordinating a group of clients to train a machine learning model collaboratively, which can be easily embedded into Unmanned Aerial Vehicle (UAV) swarms. Compared with the terrestrial wireless networks, the UAV swarm faces more precarious communication conditions, rendering synchronous aggregation no longer tenable. Additionally, the data collected from UAVs tend to be heterogeneous due to different deployment regions or requirements. To overcome these restrictions, this paper has proposed a novel two-stage Asynchronous Federated Learning scheme for the UAV swarm. Initially, the convergence property of both convex and non-convex models trained by the proposed scheme is analyzed. In the pre-training stage, we modeled the learning process as a cooperative game with demonstrated monotonicity and submodularity. Furthermore, the Shapley Value is imported to quantify data values of UAVs, and the upper bound of its estimation error rate is derived. In the training stage, a new concept named Network Age of Updates (AoU) is proposed to address the fairness issue, quantifying the model’s generalization capability with data value consideration, and a sequential UAV selection scheduling is performed through the AoU minimization by Whittle Index method. Finally, the system performance is validated through both theoretical analysis and simulations.

Sampling and interpolation of continuous graph signals have been extensively studied, in order to reconstruct or estimate the entire graph signal from the signal values on a subset of vertices. Whereas in a lot of real-world scenarios, only the signs of signals are available. For example, a rating system may only provide simple options such as "like" or "dislike". We are interested in whether it is possible to recover the original signal from such coarse information. In this paper, the reconstruction of bandlimited graph signals based on sign measurements is discussed and a greedy sampling strategy is proposed. The simulation experiments are presented, and the greedy sampling algorithm is compared with the random sampling algorithm, which verifies the feasibility of the proposed approach.

Packet routing is one of the fundamental problems in computer networks in which a router determines the next-hop of each packet in the queue to get it as quickly as possible to its destination. Reinforcement learning (RL) has been introduced to design autonomous packet routing policies with local information of stochastic packet arrival and service. However, the curse of dimensionality of RL prohibits the more comprehensive representation of dynamic network states, thus limiting its potential benefit. In this article, we propose a novel packet routing framework based onmultiagentdeep RL (DRL) in which each router possess anindependentlong short term memory (LSTM) recurrent neural network (RNN) for training and decision making in afully distributedenvironment. The LSTM RNN extracts routing features from rich information regarding backlogged packets and past actions, and effectively approximates the value function of Q-learning. We further allow each route to communicate periodically with direct neighbors so that a broader view of network state can be incorporated. The experimental results manifest that our multiagent DRL policy can strike the delicate balance between congestion-aware and shortest routes, and significantly reduce the packet delivery time in general network topologies compared with its counterparts.

Capturing complex high-order interactions among data is an important task in many scenarios. A common way to model high-order interactions is to use hypergraphs whose topology can be mathematically represented by tensors. Existing methods use a fixed-order tensor to describe the topology of the whole hypergraph, which ignores the divergence of different-order interactions. In this work, we take this divergence into consideration, and propose a multi-order hypergraph Laplacian and the corresponding total variation. Taking this total variation as a regularization term, we can utilize the topology information contained by it to smooth the hypergraph signal. This can help distinguish different-order interactions and represent high-order interactions accurately.

The problem of anomaly detection among multiple processes is considered within the framework of sequential design of experiments. The objective is an active inference strategy consisting of a selection rule governing which process to probe at each time, a stopping rule on when to terminate the detection, and a decision rule on the final detection outcome. The performance measure is the Bayes risk that takes into account not only sample complexity and detection errors, but also costs associated with switching across processes. While the problem is a partially observable Markov decision process to which optimal solutions are generally intractable, a low-complexity deterministic policy is shown to be asymptotically optimal and offer significant performance improvement over existing methods in the finite regime.

This paper investigates the active sampling for estimation of approximately bandlimited graph signals. With the assistance of a graph filter, an approximately bandlimited graph signal can be formulated by a Gaussian random field over the graph. In contrast to offline sampling set design methods which usually rely on accurate prior knowledge about the model, unknown parameters in signal and noise distribution are allowed in the proposed active sampling algorithm. The active sampling process is divided into two alternating stages: unknown parameters are first estimated by Expectation Maximization (EM), with which the next node to sample is selected based on historical observations according to predictive uncertainty. Validated by simulations compared with related approaches, the proposed algorithm can reduce the sample size to reach a certain estimation accuracy.

Packet routing is one of the fundamental problems in computer networks in which a router determines the next-hop of each packet in the queue to get it as quickly as possible to its destination. Reinforcement learning has been introduced to design the autonomous packet routing policy namely Q-routing only using local information available to each router. However, the curse of dimensionality of Q-routing prohibits the more comprehensive representation of dynamic network states, thus limiting the potential benefit of reinforcement learning. Inspired by recent success of deep reinforcement learning (DRL), we embed deep neural networks in multi-agent Q-routing. Each router possesses an independent neural network that is trained without communicating with its neighbors and makes decision locally. Two multi-agent DRL-enabled routing algorithms are proposed: one simply replaces Q-table of vanilla Q-routing by a deep neural network, and the other further employs extra information including the past actions and the destinations of non-head of line packets. Our simulation manifests that the direct substitution of Q-table by a deep neural network may not yield minimal delivery delays because the neural network does not learn more from the same input. When more information is utilized, adaptive routing policy can converge and significantly reduce the packet delivery time.

Wireless communication system incorporating unmanned aerial vehicles (UAVs) has gained much popularity recently, especially in video transmission application. This paper investigates the base station (BS) selection scheme for cellular-connected UAVs that possess the function of video collection and streaming to BS for online decision in remote processing center. We aim to minimize the expected access latency while the throughput requirement is satisfied. To this end, a sequential BS selection scheme is proposed by designing an optimal transmission rate threshold for each candidate BS. Due to the mission-driven nature, the access rate varies as the UAV moves, so an effective average transmission rate rather than instantaneous transmission rate is considered. A recursive algorithm is proposed to obtain the rate thresholds which can be used to guide whether UAV should stop or continue measuring the links of the remaining candidate BSs. The proof of the optimality of this algorithm is given. Simulation results validate the effectiveness of the proposed scheme on access latency performance compared with conventional throughput-oriented scheme.

We consider a device-to-device (D2D) underlaid cellular network, where each cellular channel can be shared by several D2D pairs and only one channel can be allocated to each D2D pair. We try to maximize the sum rate of D2D pairs while limiting the interference to cellular links. Due to the lack of global information in large scale networks, resource allocation is hard to be implemented in a centralized way. Therefore, we design a novel distributed resource allocation scheme which is based on local information and requires little coordination and communication between D2D pairs. Specifically, we decompose the original problem into two cascaded subproblems, namely channel allocation and power control. The cascaded structure of our scheme enables us to cope with them respectively. Then a two-stage algorithm is proposed. In the first stage, we model the channel allocation problem as a many-to-one matching with externalities and try to find a strongly swap-stable matching. In the second stage, we adopt a pricing mechanism and develop an iterative two-step algorithm to solve the power control problem.

In device-to-device (D2D) underlaid cellular networks, several D2D pairs can share one channel to improve the system throughput. Most existing works allow D2D pairs to reuse all the channels, which may incur complicated interference management. Therefore, each D2D pair is limited to reuse at most one channel to reduce overhead. We aim to maximize the throughput of D2D pairs while suppressing the interference to cellular links. However, the optimization problem is an intractable mixed integer non-linear programming (MINLP) problem. Meanwhile, as network size increases, acquiring the global channel state information (CSI) is expensive even impossible in practice. Therefore, we propose a novel local CSI-based distributed channel-power allocation scheme. The base station (BS) broadcasts a control signal to indicate its received interference from D2D pairs. Upon this signal, each D2D pair executes channel selection and power update independently and iteratively. Specifically, the channel allocation problem is formulated as a many-to-one matching game with externalities. The power control problem is modeled as a Stackelberg game. We prove the existence of two-sided swap-stable matching and show that the outcomes of the Stackelberg game are locally optimal under some mild conditions. Simulation results show that our scheme is efficient with low overhead.

Unraveling quality of experience (QoE) of video streaming is very challenging in bandwidth shared wireless networks. It is unclear how QoE metrics such as starvation probability and buffering time interact with dynamics of streaming traffic load. In this paper, we collect view records from one of the largest streaming providers in China over two weeks and perform an in-depth measurement study on flow arrival and viewing time that shed light on the real traffic pattern. Our most important observation is that the viewing time of streaming users fits a hyper-exponential distribution quite well. This implies that all the views can be categorized into two classes, short and long views with separated time scales. We then map the measured traffic pattern to bandwidth shared cellular networks and propose an analytical framework to compute the closed-form starvation probability on the basis of ordinary differential equations (ODEs). Our framework can be naturally extended to investigate practical issues including the progressive downloading and the finite video duration. Extensive trace-driven simulations validate the accuracy of our models. Our study reveals that the starvation metrics of the short and long views possess different sensitivities to the scheduling priority at base station (BS). Hence, a better QoE tradeoff between the short and long views has a potential to be leveraged by offering them different scheduling weights. The flow differentiation involves tremendous technical and non-technical challenges because video content is owned by content providers but not the network operators and the viewing time of each session is unknown beforehand. To overcome these difficulties, we propose an online Bayesian approach to infer the viewing time of each incoming flow with the “least” information from content providers.

Distributed filtering in network is a fundamental problem in the field of network signal processing. Each node estimates or tracks some unknown state relying on the private observation and the fusion information from the network. Network fusion is generally a way of interaction over network, by which nodes can learn from each other and make decision mutually. Unlike conventional methods, we construct a distributed filter using Bayesian network game as a fusion tool, where all the nodes exchange their best strategies instead of exchanging local estimators. The proposed algorithm is a coalition of signal processing and game theory in network, which can be extended to more general signal processing and decision making models.

In this paper, a low-complexity sparse channel estimation scheme is proposed for massive three dimensional multi-input multi-output (3D-MIMO) systems. In outdoor propagation environment, 3D-MIMO channels exhibit joint sparseness in both temporal and angular domains, sharing the common support in delay domain. By taking prior knowledge of the structured sparseness, the proposed heuristic channel estimation method can greatly reduce the complexity of channel estimation, and achieve a near optimal performance. Simulation results verify the effectiveness of the proposed algorithm.

This paper investigates simultaneous wireless information and power transfer (SWIPT) in K-user multiple-input multiple-output (MIMO) interference channels. In particular, the power splitting (PS) technique is leveraged at each receiver to divide the received signal into two flows, for information decoding (ID) and energy harvesting (EH), respectively. As a whole system, our objective is to minimize the total transmit power of all transmitters by jointly designing transmit beamformers, power splitters, and receive filters, subject to the signal-to-interference-plus-noise ratio (SINR) constraint for ID and the harvested power constraint for EH at each receiver. Due to the coupling nature of all variables, the formulated joint transceiver design problem is nonconvex, and has not yet been well addressed in the literature. In this paper, we first propose a semidefinite relaxation-based alternating optimization (SDRAO) solution to approach the optimal solution of the problem. Then, we semidecouple the joint optimization by the derived diversity interference alignment (DIA) technique, and obtain a solution of lower complexity. Finally, a closed-form solution is further developed relying on the transmitter-side zero-forcing (TZF), which can be implemented in a distributed manner, with the lowest computational complexity and CSI exchanging overhead.

In today WSNs, sensor nodes are able to obtain energy from ambient with energy-harvesting components. However, the energy consumption are diverse across these nodes due to functional or geographical variation, which may lead to potential energy imbalance in network. In virtue of recent wireless power transfer (WPT) technology, the imbalance can be alleviated if all sensor nodes share energy with each other. In order to achieve the maximum energetically sustainable workload, we design an energy cooperation strategy in network by WPT, named energy routing, which should be jointly optimized with data routing simultaneously. An iterative distributed algorithm is developed to achieve the optimal data routing and energy routing solutions, where all sensors only need to exchange local information with neighbors. Simulation results show that the proposed algorithm can achieve higher workload than algorithms without energy cooperation.

This paper investigates the beamforming for three- dimensional multiple input multiple output (3D-MIMO) systems with inaccurate channel state information (CSI). From the view of angle-domain, the 3D-MIMO channel is sparse on the high 3D resolution provided by planar antenna array with large number of antenna elements at the base station (BS) in 3D-MIMO systems. Prior knowledge of sparsity is not only beneficial to channel estimation in literature, but also implies more efficient beamforming with inaccurate CSI at transmitters as discussed in this paper. We prove that the optimal beamforming vector is correspondingly sparse in angle-domain with a sparse channel. Therefore, we add the ℓ1-norm penalty to the beamforming vector in optimization design in angle-domain, which can fight against the perturbation due to inaccurate CSI. Technically, the problem is reformulated as a second order cone program (SOCP) form that can be solved efficiently. Simulation results demonstrate that the proposed beamforming method can achieve considerable system sum-rate improvement with high CSI error.

Unraveling quality of experience (QoE) of video streaming is very challenging in bandwidth shared wireless networks. It is unclear how QoE metrics such as buffering time and starvation behavior interact with dynamics of streaming traffic load. In this paper, we collect view records from one of the largest streaming providers in China over two weeks and perform an in-depth measurement study on flow arrival and viewing time that shed light on realistic streaming traffic pattern. Our most important observation is that the viewing time of streaming users fits a hyper-exponential distribution quite well. This implies that all the videos can be categorized into two classes, short and long viewing time with separated time scales. We then map the traffic pattern of large-scale measurement to bandwidth sharing cellular networks. We propose two models to compute the close-form starvation probability and mean sojourn time on the basis of ordinary differential equations (ODEs). Extensive trace-driven simulations validate their accuracy. The proposed models precisely capture how the QoE metrics of video streaming in each class are influenced by the scheduling algorithms at a base station.

This paper focuses on the simultaneous wireless information and power transfer (SWIPT) in coordinated multi-cell networks. In particular, the considered system has a location-aware feature, i.e., the conventional coordinated multi-point (CoMP) information transmission is available to the cell-edge users; meanwhile, the emerging wireless power transmission is available to the cell-center users. Subject to the SINR constraints for information decoding at edge CoMP users and the received power constraints for energy storage at center users, our objective is to minimize the total transmit power of all coordinated beamformers. Although the celebrated semi-definite relaxation (SDR) technique can be applied to obtain a centralized optimal solution, we prefer to decompose the original problem and derive a decentralized closed-form solution to decrease the cost of inter- cell cooperation. With this low-complexity solution, some important insights are further presented on the benefit of the location-aware SWIPT setup and the potential user scheduling mechanism. Simulation results validate our derivation and show that the decentralized solution can asymptotically approach the centralized optimal solution.

Many signal processing problems in wireless sensor networks can be solved by graph filtering techniques. Finite impulse response (FIR) graph filters (GFs) have received more attention in the literature because they enable distributed computation by the sensors. However, FIR GFs are limited in their ability to represent the global information of the network. This letter proposes a family of GFs with infinite impulse response (IIR) and provides algorithms for their distributed realization in wireless sensor networks. IIR GFs bring more flexibility to GF designers, as they can be designed and realized even when the graph spectrum is unknown. Numerical results show that IIR GFs are more accurate in approximating ideal GFs and more robust against network variation than FIR GFs.

Channel prediction is an appealing technique to mitigate the performance degradation due to the inevitable feedback delay of the channel state information (CSI) in modern wireless systems. We first propose a general MIMO-OFDM channel prediction framework, which exploits both the spatial and temporal correlations among antennas. Then we derive two predictors which select data for auto-regressive (AR) predictors in different ways based on the proposed framework. The first predictor chooses the data set via minimizing the mean square error (MSE) of prediction model. The second predictor chooses the data in a heuristic way, which aims to reduce the computational complexity. Our algorithms can be applied to improve the precoding performance in multi-user MIMO-OFDM systems. Simulation results show that the proposed methods can overcome the feedback delay effectively, even when the channel changes rapidly.

Clock synchronization is an essential service for various applications in wireless sensor networks (WSNs). Most of available synchronization protocols are based on CSMA MAC scheme. They are energy inefficient in dense sensor networks. Despite many power-saving TDMA MAC protocols have been proposed, few of them are qualified to flood synchronization messages. The novel one here adopts a carefully designed time division flooding scheme, which divides time into slots and assigns each node a unique one for transmitting synchronization packets. In this way, the active duration of each node can be dramatically reduced. The new slot allocation policy guarantees active periods of nodes along the forwarding paths be scheduled successively, so synchronization packets can be spread out in a short time. Besides, a special clock frequency skew compensation technique is employed to prolong the re-synchronization interval. Simulation shows that our protocol achieves high energy efficiency and fast convergence speed.