Given the emerging needs of artificial intelligence in both commercial and consumer uses, the rapid increase of energy consumption has become a major concern for data centers. Workload scheduling plays a crucial role in reducing the energy costs associated with large compute-intensive workloads in data centers. However, there has not been much research on reducing the energy costs for GPU-intensive workloads specifically. In this paper, we proposed two thermal-ware scheduling algorithms (GPUThermalAwareMinTemp and GPUThermalAware) for GPU-intensive workloads. Through the integration of machine learning models into a modified version of the GPUCloudSimPlus simulator, we compare the performance of our algorithms with three baseline algorithms by simulating cloud computing environments using a subset of real-world workload traces from Alibaba PAI. Our experimental results show that, compared with the worst-case algorithms, the GPU thermal-aware scheduling algorithms achieved 4.72% lower average host minimum temperatures and saved total energy by 1.17%.
To overcome signal blockage and extend coverage to both indoor and outdoor users, simultaneous transmitting and reflecting reconfigurable intelligent surfaces (STAR-RIS) are employed. In addition, Non-Orthogonal Multiple Access (NOMA) is integrated to improve spectral efficiency and support multipleuser access. Given this setup, it is essential to investigate the Physical Layer Security (PLS) performance of STAR-RIS-assisted NOMA systems, particularly with multiple eavesdroppers. To evaluate the signals' interception risks, we derive closed-form expressions for the Secrecy Outage Probability (SOP) and the Intercept Probability (IP) for both system setups at the near users and colluding/non-colluding eavesdroppers.
Satellite communication networks - including Low Earth Orbit (LEO) and Geostationary Earth Orbit (GEO) systems - are becoming increasingly vital for global connectivity, particularly in remote and underserved areas. Despite their growing importance, individual satellite links often suffer from high latency, limited bandwidth, and fluctuating reliability, which can significantly degrade application performance. Multipath Quick UDP Internet Connection (MPQUIC) has emerged as a promising solution for aggregating multiple satellite paths and improving network utilization. However, two major challenges limit the effectiveness of existing MPQUIC schedulers: the heterogeneity of link characteristics across satellite types and the periodicity of LEO link behavior. To address these issues, this paper proposes Logr alpha, a novel MPQUIC scheduler that integrates Deep Reinforcement Learning (DRL) to dynamically optimize data distribution and manage adaptation speed. Logr alpha intelligently adjusts path usage to account for link diversity and adaptively adjusts the growth rate of the congestion window to respond quickly to periodic variations in LEO links. Experimental results show that Logr alpha significantly enhances adaptation speed and reduces file transfer time compared to traditional scheduling strategies.
A neural architecture for user-centric vehicular caching is introduced, termed as CacheFormer. It leverages autoformer for cache content prediction that positions vehicles as autonomous micro-servers that operate independently within personal area networks. The architecture employs multi-horizon prediction with five-minute slot granularity. It enables lookahead through decomposition of application patterns into trend and seasonal components. Integration of sparse auto-correlation mechanisms yields in reduced complexity while maintaining prediction accuracy. Further, an age-aware eviction mechanism combines forecasted demand with cache staleness metrics, which eliminates separate prediction and eviction modules. Experimental evaluation on 3,463 application sessions demonstrates convergence within 50 epochs in 27.5 minutes on consumer hardware, at high hit rates post-convergence across all popularity quartiles. Moreover, bimodal correlation distribution reveals successful temporal pattern extraction for 75% of applications. Results validate that adopting autoformer improves cache efficiency while preserving data sovereignty and eliminating dependencies on external cloud infrastructure during user mobility.
With the increasing adoption of Unmanned Aerial Vehicles (UAVs) and their technological advancements, particularly in terms of autonomy and computational capabilities, their role as active component in wireless networks could be groundbreaking. Such potential is especially relevant in constrained environments that lack proper telecommunication infrastructure or adequate network coverage. These contexts are where the Internet of Things (IoT) has emerged as a vital technology, capable of facilitating sensing operations and data retrieval by leveraging communication protocols like LoRa for long-range, low-energy transmissions. In this paper, we introduce the Traveland-Offload Problem (TOP), a new combinatorial optimization problem, formulated both as a mixed-integer nonlinear programming (MINLP) model and as a mixed-integer linear programming (MILP) model. Its primary objective is to minimize the energy consumption associated with UAV operations and data offloading tasks. Our formulation, which incorporates realistic models for energy consumption and communication, aims to optimally determine the UAV’s trajectory, hovering points and durations, as well as critical transmission parameters of LoRa-enabled sensors, while ensuring compliance with the transmission policies defined by the LoRaWAN network protocol.
Optical wireless communication (OWC) is increasingly explored as a high-capacity, energy-efficient alternative for terrestrial and non-terrestrial links. To address challenges posed by mobility-induced Doppler shifts, orthogonal time frequency space (OTFS) modulation offers enhanced spectral efficiency and robustness, making it well-suited for dynamic OWC scenarios. This paper presents one of the first real-time software-defined radio (SDR)-based implementations of optical OTFS (O-OTFS) tailored for intensity-modulation direct detection (IM/DD) optical systems. Two modulation schemes—direct current-biased O-OTFS (DCO-OTFS) and asymmetrically clipped O-OTFS (ACO-OTFS)—were designed in MATLAB Simulink and deployed on USRP B210 radios. The SDR platform enables flexible signal generation, over-the-air prototyping, and integrated optical testing. Experimental validation confirms the feasibility of both schemes, achieving bit error ratios (BERs) on the order of 10-4, suitable for integration with standard forward error correction (FEC) schemes, demonstrating the viability of SDR-based prototyping for next-generation OWC links.
Large Language Models (LLMs) are rapidly being embedded in consumer and service robots, enabling richer human-robot interaction, multimodal reasoning, and language-driven autonomy. However, the computational and lifecycle costs of training, inference, and continuous upgrade cycles raise urgent digital sustainability concerns: energy consumption, network dependency, privacy exposure, and hardware obsolescence. In this conceptual paper, we introduce the Sustainable Language-Driven Autonomy Framework (SLAF), a modular architecture and set of operational policies that align multimodal LLM reasoning with sustainability goals. SLAF decomposes the intelligence stack into Perception & Preprocessing, Local Cognition (Edge), High-level Reasoning (LLM), and Control & Execution layers, mediated by an Adapter responsible for compact semantic encoding, adaptive triggers, caching, and energy budgets. We propose quantitative primitives and trade-off models (e.g., energy-per-inference Einf, calls-per-mission Ncalls, mission energy Emission) and an evaluation protocol to make sustainability claims comparable and auditable. Finally, we map how SLAF addresses four research questions on zero-shot generalization, energy-efficient architectures, software-first lifespan extension, and cloud/on-device trade-offs. We conclude with a roadmap for empirical validation, lifecycle analysis, and user-centered studies to operationalize sustainable, language-enabled robotics.
This paper presents a heuristic for simplifying resource allocation in wireless access systems using the concept of comparative advantage. It reduces allocation complexity in multi-cell networks by halving the dimensionality of the search space before optimization. Simulations and experiments confirm its effectiveness, highlighting its potential for practical deployment in next-generation wireless access networks.
The advent of Internet of Things (IoT) ushers in a significant potential to integrate individuals, devices, and data, leading to a profound change in our professional and social environments. The small, resource-constrained IoT devices are usually deployed to collect various types of critical data in remote or unmonitored locations. Due to extensive interconnectivity, limited resources, and inadequate security design, IoT systems are vulnerable to communication-specific cyber threats, which aim to disrupt operations, steal sensitive information, or cause damage. To resolve the security concerns in IoT communications, many recent efforts have been devoted to designing authenticated key agreement protocols for IoT systems. However, not only most of the existing solutions fail to adopt cost-effective techniques for resource-limited IoT devices, but also they ignore the differentiation among various data types in the established session keys. A few approaches use traditional physical unclonable functions (PUFs) to address resource concerns, yet they introduce new security issues into IoT systems. Once the PUF cryptographic information is compromised by machine learning attacks, the entire authentication framework collapses. Therefore, in this paper we propose an authenticated key agreement protocol for device-to-gateway communication in IoT systems based on Chebyshev polynomial and probability-based PUF. We examine the proposed protocol’s security features through formal security validation. We also conduct performance evaluation through a simulation-oriented study, and the results clearly prove that the proposed protocol offers superior security and privacy, while maintaining low computational overhead.
Power Line Communication (PLC) systems are facing increasing security threats as adversaries leverage low-cost Software-Defined Radios (SDRs) to launch physical-layer attacks, e.g., jamming and Radio Frequency Fingerprinting (RFF), for communication disruption and unauthorized device tracking, respectively. This paper investigates the dual role of Radio Frequency (RF) wireless jamming for PLC environments, through two distinct scenarios: (i) friendly RF jamming for privacy preservation of (cabled) PLC devices against unauthorized RFF, and (ii) adversarial RF jamming to degrade the performance of legitimate RFF-based authentication systems. We conducted various systematic experiments using nine USRP X310 SDRs connected to actual PLC couplers exchanging signals modulated according to the Binary-Phase Shift Keying modulation scheme to analyze the behavior of RFF in PLC scenarios under different RF jamming levels. Our results demonstrate, for the first time, that strategic RF jamming effectively obscures device fingerprints in cabled PLC communications while maintaining communication quality, with bit error rates remaining acceptable across most configurations. We also demonstrate that device identification accuracy degrades significantly as the jamming intensity increases. Our findings establish fundamental trade-offs between privacy protection and authentication reliability, providing insights for the design of robust PLC systems.
Energy harvesting using communication radio waves transmitted from base stations is a promising technology for realizing battery-less IoT (B-IoT) devices demanded in the beyond 5G era. However, when downlink traffic is small, the power of radio waves from the base station is low and the generated power for B-IoT devices decreases, which hinders their widespread application. To solve the above problem, we propose a downlink resource reservation method based on resonant frequency to maximize the efficiency of radio frequency to direct current (RF-DC) converter converting radio waves to electric power. In our proposed method, frequency resources with high conversion efficiency are reserved in the wireless resources divided by the time and frequency axis. The reserved wireless resources are utilized for power supply to B-IoT devices in the downlink from the base station. Therefore, even when the downlink traffic is small, the power of radio waves from the base station is stably high, and the power conversion efficiency in the B-IoT device is also high, so that more power can be obtained. In this paper, we measured the resonant frequency of the RF-DC converter and evaluated the proposed method using the resonant frequency and universal software radio peripheral. The experimental results show that when the frequency band of 30 MHz is reserved, the voltage generated by the proposed method is about 101% higher than that by the method reserving from behind the radio resources.
The persistent digital divide, characterized by un-equal access to affordable and reliable mobile connectivity, remains a critical challenge for inclusive global development. The flow of network resources remains a complex task due to multi-vendor policies in beyond 5G networks (B5G) and sparse user distributions in remote areas. Addressing this issue, this paper presents a cost-aware, game-theoretic framework that leverages software-defined network slicing (SDNS) to minimize deployment cost while maintaining service quality across distributed and sparsely connected user entities (UEs). The problem considers multiple mobile infrastructure providers (MIPs) offering network slices at different cost levels and mobile virtual network operators (MVNOs) serving end-users seeking optimal connectivity at reasonable prices. The slice allocation and pricing formulation is cast as a mixed-integer nonlinear optimization problem (MILP) that is NP-hard due to the combinatorial coupling of users, slices, and link capacities. We decompose the problem into tractable subgames using a non-cooperative game-theoretic approach that converges to a near-optimal Nash equilibrium. The equilibrium ensures fair profit distribution among MIPs while providing affordable access to users. Simulation results over sparse network topologies demonstrate improved throughput and SINR characteristics with cost savings up to 25% compared to random allocation. Profitability indices show balanced revenue across providers, validating the efficiency of the proposed distributed algorithm in bridging the digital divide.
Digital Twins (DTs) are increasingly adopted in Internet of Things (IoT) systems to provide real-time virtual representations of physical assets. Their growing interconnectivity, however, exposes them to cybersecurity threats that can compromise fidelity and reliability. This work-in-progress study evaluates how representative attacks—Denial of Service, packet loss, and message manipulation—affect DT operational metrics: timeliness, reliability, availability, and consistency. Using a simulation framework based on NS-3, MQTT, and Eclipse Ditto, we emulate realistic DT environments and quantify the resulting degradations. Results show that cyber-attacks leave measurable deviations in these metrics, and that a revised Overall Digital Twin Entanglement (ODTE) score calibrated on normal operation effectively amplifies such anomalies. Metric-based observability thus emerges as a lightweight, non-intrusive approach for early detection and resilience assessment in DT-enabled infrastructures.
This paper surveys the integration of Digital Twins (DT) and Federated Learning (FL) in Industrial IoT (IIoT), highlighting opportunities for real-time monitoring, predictive analytics and distributed intelligence. We review state-of-the-art approaches, identify technological and methodological challenges and discuss how DT and FL can be jointly leveraged to support secure, resilient and adaptive industrial operations. Beyond the classical focus on assets and processes, we extend the analysis to Vulnerable Road Users (VRUs) and show how this concept can be transposed into industrial contexts, where workers and operators act as VRUs inside plants, warehouses, and construction sites. In this perspective, DTs combined with FL can provide simulation-driven insights for worker-centric safety, enabling risk prediction, proactive protection and safer coordination between humans, machines and autonomous systems.
Low Earth Orbit (LEO) satellite constellation systems are expected to be integrated with terrestrial networks for future communication services over wide areas. For real-time applications, direct data relay between users is effective to minimize the propagation delay via LEO satellites, where multi-beam control is an essential function in LEO satellites to efficiently allocate beam resources over the areas. However, no previous works considered the satellite’s buffer limitations and the differences in throughput between uplink (UL) and downlink (DL) which caused queuing delays. This paper proposes a coordinated UL/DL multi-beam scheduling to minimize the end-to-end latency of data relay by considering the available buffer space of the satellite and variations in UL/DL throughput due to satellite movement. An optimization problem is formulated for the scheduling design to minimize the end-to-end latency. We demonstrate with simulation results that the proposed scheduling method reduces the latency by 16.0%, compared to the conventional methods.
3GPP standardized conditional handover (CHO) in Release 17 as a mechanism to enhance mobility robustness in non-terrestrial networks (NTNs). In addition to the traditional terrestrial NR triggering events (e.g., Event A3), two new triggers, location-based and time-based, were introduced. These enhancements address the limitations of relying solely on signal strength, which can be insufficient for distinguishing between satellites at similar altitudes due to comparable path-loss characteristics. In this paper, a time-to-exit (TTE) criterion, integrated with a location-based trigger, is proposed to select candidate cells for which the network estimates how long the D1 condition will remain valid after handover. A candidate cell is admitted if (i) it has the longest D1-based TTE and (ii) its signal quality exceeds a configured threshold; otherwise, it is discarded, with ties broken based on signal strength. This design arms only targets that are both strong enough and stable long enough to finish the handover, preventing late or unstable executions. Simulation results show that the proposed TTE-aware CHO significantly reduces handover failures and ping-pong events, and increases post-handover time-of-stay, compared with a baseline location-based CHO.
In edge computing environments, challenges in utilizing FPGAs as computational resources in edge nodes include the separation of expertise among FPGA designers, network engineers, and IoT engineers, as well as issues of logic reusability and flexible configuration. A GUI-based framework using Node-RED is proposed, leveraging the reconfigurability of FPGAs and dynamic network configuration to enable collaboration among engineers from different fields.The proposed framework utilizes the Processing System (PS) and Programmable Logic (PL) of SoC FPGAs, allowing visual operation of FPGA nodes on Node-RED. When FPGA nodes are connected consecutively, direct data transfer between FPGAs is possible without passing through Node-RED, minimizing communication latency and overhead. Furthermore, by dynamically configuring network settings instead of hard-coding them into the logic, the same logic design can be reused in different network environments, reducing development costs and enabling rapid service deployment.In evaluation experiments, communication latency between Node-RED and FPGAs was measured in both two-point and three-point communication scenarios. The results confirmed that the additional latency introduced by the proposed framework is minimal and does not compromise the processing speed advantage of hardware accelerators. Additionally, resource evaluation showed that sufficient resources can be secured, indicating the possibility of implementing additional applications.With this framework, FPGA designers can focus on logic design without considering network or communication partners, while IoT engineers can use FPGAs as applications without knowledge of FPGA or coding. This improves the practicality and reusability of FPGAs in edge computing environments and demonstrates the potential for more engineers to participate in FPGA-based application development.
Recent advances in multimodal learning have enabled powerful Vision-Language-Action (VLA) systems for robotic reasoning and control. However, most existing approaches rely on Transformer backbones, which face scalability and efficiency bottlenecks for long sequences. This work introduces MambaVLA, a scalable VLA framework built on the Mamba state space architecture for efficient sequence modeling. The framework integrates the Eagle visual encoder and Qwen-7B-Chat-Int4 language model to achieve fine-grained multimodal fusion with linear-time complexity. A diffusion flow matching module further aligns visual-language embeddings with continuous action trajectories, enabling smooth and precise control. Extensive evaluations on standard VLA benchmarks demonstrate that MambaVLA matches or surpasses Transformer based models while offering substantially lower computational cost and faster inference. These results highlight the potential of state space modeling and flow-based action generation for compact, scalable, and deployable embodied intelligence systems. https://sainavaneet.github.io/MambaVLA.gihub.io/
The extensive value and importance of smart grids are evident in their transformative impact on sustainable urban development. However, the swift progression of quantum computing has made the security and privacy of smart grid communications a pressing and vital issue. Although considerable studies have been carried out on authentication and key agreement within smart grids, the predominant body of solutions either come with substantial computation, communication, and storage overheads, or are primarily single-layer schemes. More importantly, they are not designed to withstand advanced quantum attacks. In this paper, we propose a novel quantum-safe and cross-layer authentication and key agreement protocol, named QCLaka, for smart grid communications that overcomes the shortcomings of existing approaches and offers advanced security and functionality features. The proposed QCLaka protocol integrates lattice-based cryptography with probability physical unclonable function (Prob-PUF) to enable the local gateway and the smart meter to mutually authenticate each other and establish a secure session key. The security of the proposed QCLaka protocol is assessed through formal security verification to highlight its safety in adversarial environments and its ability to withstand both well-known and advanced cyberattacks. Additionally, the experimental evaluation shows that the proposed QCLaka protocol outperforms the benchmark schemes regarding the performance in security and efficiency.
