Air and maritime surveillance impose stringent requirements on the underlying backbone communication network in terms of availability, latency, traffic differentiation, and operational sovereignty, especially where fiber-optic deployment is impractical. Existing studies focus on individual link design or theoretical resilience mechanisms, leaving the system-level integration into a nationwide operational scenario underexplored. This article presents the design and quantitative validation of a 42-link, 40-site microwave backbone in the defense-reserved 4-GHz band, with frequency division duplex (FDD) channelization compliant with ITU-R F.1099-5, Carrier Ethernet transport, and QinQ-based service segregation for radar, very/ultra high frequency (V/UHF) voice, and command-and-control traffic. The architecture is validated through radio-electrical simulation over real terrain covering peninsular Spain and the Balearic Islands, spanning 4165 km of aggregated link distance with link spans from 26 to 143 km. All 42 links exceed the 32-quadrature amplitude modulation (QAM) threshold by an average margin of 33.9 dB, enabling adaptive modulation up to 256-QAM (200 Mbps per link), with an end-to-end availability of 99.98% across three disjoint paths between command centers. A comparative assessment against fiber, geostationary Earth orbit (GEO), low-Earth-orbit sattelite, and hybrid alternatives confirms the microwave backbone as the most suitable solution for the considered scenario.
This article investigates fully distributed event-triggered prescribed-time consensus for a class of second-order nonlinear multiagent systems with disturbances. A tunable prescribed-time coefficient function (TPTCF) is introduced to generalize existing prescribed-time functions and provide additional flexibility for moderating the growth of the time-varying control gain. Based on the TPTCF, two event-triggered control strategies are developed to reduce unnecessary updates and improve resource utilization. The first strategy achieves prescribed-time consensus under a constrained leader velocity through a fully distributed protocol. The second strategy incorporates an event-triggered prescribed-time observer and removes the constraint on the leader. Moreover, the two strategies involve three dynamic threshold functions, two of which are used in the second strategy, to meet prescribed-time requirements. Fuzzy logic systems and adaptive techniques are employed to handle unknown nonlinearities and external disturbances. Finally, two numerical examples verify the effectiveness of the proposed methods.
Leveraging alternative energy sources and load flexibility in interconnected infrastructures can enhance the recovery of power distribution networks (PDNs) during low-probability, high-impact events. This study proposes an emergency coordination framework between PDNs and water distribution networks (WDNs) that co-optimizes water pumps as flexible loads and pumps-as-turbines (PATs), as distributed energy sources through strategic pressure management. The framework consists of three steps: 1) PDN shares the restoration start time and emergency duration with WDN; 2) WDN schedules pumps and PATs by solving two pressure-minimizing hydraulic optimization subproblems and reports the resulting power profiles to the PDN; 3) PDN solves a mixed integer linear programming to maximize prioritized load restoration using the provided information, microgrids formation, and tie-line operation, and then signals whether to implement the WDN schedule. Case studies on the IEEE 33-bus PDN integrated with the NET3 WDN demonstrate the framework’s effectiveness. Across scenarios with varying line outages and emergency durations, coordinated operation increases average restored active loads from 33.31% (without coordination) to 39.58% (with coordination), and to 40.03% when a PAT supplies 10.60 kW on average. Correspondingly, WDN pressure satisfaction adjusts from 100% to 96.77% and 94.76%, highlighting the importance of explicit resilience tradeoff analysis.
Cascading failures in power systems are complex, high-impact phenomena that can result in significant financial and societal impacts. Accurately predicting the severity of these cascades and the postcascade status of transmission lines early in the event can greatly aid in risk assessment and the formulation of effective mitigation strategies. Graph neural network (GNN)-based models are emerging as powerful tools for such predictive analyses in power systems due to their ability to capture topological relationships among system measurements. In this work, various GNN-based models are proposed to improve the prediction of transmission line statuses and overall cascade size using the initial cascade data. Performance evaluations on the IEEE 118-bus and 300-bus systems demonstrate that GNN-based models outperform the traditional machine learning models that do not incorporate topological information. In addition, the sensitivity of the proposed GNN-based models to noise in the initial cascade data is evaluated, highlighting the critical importance of accurate initial trigger information.
System of systems (SoS) warfare has become a predominant paradigm in contemporary military operations. Within this framework, heterogeneous equipment is organized into integrated equipment SoS to accomplish missions through complex interactions and interdependencies. Designing an effective and efficient equipment SoS requires a rigorous identification of its capabilities and a sound assessment of their relative importance. However, the intricate relationships among tasks and capabilities make capability analysis challenging. To address this issue, this study proposes an integrated quality function deployment (QFD) approach that combines the analytic hierarchy process (AHP), the decision making trial and evaluation laboratory (DEMATEL) method, and the Tomada de Decisao Interativa Multicriterio (TODIM) within a Fermatean fuzzy environment. Fermatean fuzzy sets are used to represent uncertain and subjective information. Within the proposed framework, FF-AHP is employed to determine task importance, FF-DEMATEL is adopted to analyse intercapability relationships, and Fermatean fuzzy TODIM is used to derive a prioritized ranking of capabilities by jointly considering task importance, capability interdependencies, and task capability relationships. A case study on a search and rescue mission is conducted to illustrate the applicability of the proposed method. The results show that the framework can identify critical capabilities that are consistent with operational knowledge, and comparative analyses with conventional QFD, fuzzy technique for order preference by similarity to ideal solution (TOPSIS), fuzzy VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) and Fermatean fuzzy TOPSIS confirm the robustness and effectiveness of the proposed approach. Overall, the integrated Fermatean fuzzy QFD methodology provides a systematic and reliable tool for capability analysis in complex SoS warfare environments.
Social media platforms shape social media sentiment related to adolescent health topics, including obesity. The proposed framework demonstrates how large language model (LLM)-based refinement enhances contextual interpretability in sentiment analysis. The pipeline performs collection, preprocessing, account classification into five source categories, lexicon scoring, LLM refinement applied to all cleaned comments, and visualization. We have evaluated the system on a curated corpus of approximately 21 000 public comments collected from TikTok ($n \approx 20{\,}900$) and Instagram ($n \approx 238$). Statistical analyses including chi-square, Mann–Whitney $U$, and bootstrapped confidence intervals reveal platform-specific differences in sentiment polarity distributions: TikTok exhibits greater affective variability, while Instagram comments are concentrated toward moderately positive polarity. Finally, we conclude with design recommendations for platform-sensitive adolescent health communication and discuss ethical and operational considerations for deploying LLMs in public health analytics.
Probabilistic energy flow (PEF) is the foundation for the operation and planning of integrated power-gas systems (IPGS) considering multiple uncertainties from sources and loads. However, the PEF calculation based on traditional polynomial chaos expansion (PCE) incurs significantly higher computational cost in high-dimensional scenarios. Hence, an efficient PEF calculation method based on the sparse polynomial chaos expansion (SPCE) is proposed. First, a steady-state energy flow calculation model for IPGS is established considering the source-load uncertainty and the wind-solar correlation. Then, based on the principle of PCE, the bias estimation and cross-validation methods are used to choose the optimal expansion terms. Through a cyclic iteration involving the addition and deletion terms, the contribution of each expansion term is assessed to acquire the sparse polynomial. Finally, the probability distributions of each energy flow state variable of two typical IPGS are studied by the proposed SPCE method. Results indicate that the SPCE method can efficiently calculate the PEF and substantially reduce the computational cost compared with the traditional PCE and existing q-PCE methods while ensuring accuracy, particularly in the multidimensional input variables.
Efficient 3-D path planning for autonomous underwater vehicles (AUVs) in dynamic submarine environments presents a significant challenge due to complex seabed terrain, ocean currents, and obstacles. In view of the adaptability and generalization limitations of traditional methods, this article proposes the reward-adaptive prioritized experience replay (RAPER) mechanism to dynamically adjust the experience sampling priorities by evaluating the temporal-difference errors and the weights of critical reward events, including success rate, goal proximity, route length, obstacle avoidance, and ocean current utilization. Then, a deep reinforcement learning framework, the DSAC-T-RAPER algorithm, is established for underactuated AUVs by combining the RAPER mechanism and distributional soft actor–critic with three refinements (DSAC-T) algorithm. Meanwhile, the finite-step evolution and boundedness of the adaptive event weight are theoretically analyzed to guarantee the reliability and stability of the proposed algorithm. The 3-D simulation environment is constructed by considering complicated submarine seafloor topography, real ocean current data from the Copernicus Marine Environment Monitoring Service, and randomly distributed obstacles. Simulations under different scenarios demonstrate that, compared with some other algorithms, DSAC-T-RAPER achieves higher success rates and better performances.
Digital twin systems require continuous synchronization with distributed Internet of Things devices to maintain accurate representations of physical processes, but frequent updates incur significant energy and latency overhead. Conventional strategies rely on static parameters that do not adapt to time-varying dynamics, leading to inefficient resource usage or degraded fidelity. This article introduces a unified evaluation framework integrating energy consumption, reconstruction accuracy, and latency through the energy-aware twin update for networked environments metric, and proposes a drift-adaptive synchronization mechanism that dynamically adjusts update thresholds based on observed signal drift. The energy model is grounded in the Si4460 transceiver hardware of the evaluation dataset, where transmission energy dominates the energy consumed during deep-sleep operation, validating the use of a per-transmission energy proxy across all policies. Evaluation on real-world crowd-sensing traces and synthetic workloads shows that the adaptive mechanism achieves favorable energy–fidelity–freshness tradeoffs under moderate dynamics and remains competitive across static and highly dynamic operating regimes. Age of Information (AoI) analysis further shows that the adaptive policy achieves the lowest mean staleness, reducing AoI by 55% relative to periodic synchronization. These results demonstrate that drift-adaptive synchronization provides an effective and lightweight solution for energy-efficient digital twin systems.
Most existing results on nonlinear systems with unknown control directions (CDs) focus on asymptotic tracking. This article investigates the problem of adaptive predefined-time (PT) precise tracking for a class of nonlinear multiagent systems (MASs) with unknown finitely switching CDs and time-varying input delay. A novel control framework is established by introducing monotonically increasing sequences, which extends the classical Nussbaum function methodology to scenarios involving finitely switchable and completely unknown CDs via a proof by contradiction mechanism. Moreover, to compensate for the input delay, a class of delay compensation signals is introduced. By incorporating fuzzy logic systems and bounded estimation techniques, it is rigorously proven that the tracking error converges to zero within a PT, while simultaneously ensuring that full-state constraints are satisfied and all closed-loop signals are semiglobally ultimately bounded. The proposed method offers improved control performance and a broader applicability compared to existing approaches. Numerical simulations are provided to demonstrate the effectiveness of the proposed strategy.
This article develops a guidance strategy for air defense missiles to intercept maneuvering targets under defender protection. We first address the fundamental scenario where a single defender guards the target, proposing a novel heading-error barrier function based on barrier function theory to ensure that the attacker maintains a safe distance from high-maneuvering defenders. The resulting constraints are incorporated into a quadratic programming (QP) framework that minimizes lateral guidance effort while yielding a closed-form solution. Building upon this foundation, we significantly extend our method to handle the more complex case of multiple defenders by introducing a soft-minimum approach that smoothly combines multiple heading-error barrier functions. This innovative transformation reduces the multiconstrained QP problem to a single-constrained formulation while preserving safety guarantees. The proposed framework offers notable flexibility, capable of accommodating various guidance strategies including those with impact angle specifications. Comprehensive simulations validate the method's effectiveness across different engagement scenarios, demonstrating its versatility in both single-defender and multidefender environments.
This article proposes a novel distributed event-triggered compensation interval observer (ETCIO) for a general nonlinear interconnected (NI) system with disturbances in the state and output vectors, where the nonlinear function depends not only on the local state vector but also on the remote state vectors. In contrast to the existing event-triggered interval observers, which are used to estimate the state vectors of single systems (systems without interconnection between subsystems), the one in this article can be used to estimate the state vector of each subsystem of an NI system. Unlike the existing distributed event-triggered interval observers of interconnected systems, where the issue of compensation measurement for the missing output information over interevent intervals was not considered, and the nonlinear function only depends on the local state vector (which may limit its application), the one in this article can mitigate the loss of output information over interevent intervals and can be applied to a general NI system. The distributed ETCIOs are first designed. Then, the existence conditions of the proposed distributed ETCIOs are derived, and a convex optimization problem is established to minimize the bounds of the proposed ETCIOs. Third, an effective algorithm is derived to provide unknown matrices of the proposed ETCIOs. Finally, the obtained theoretical results are demonstrated by a numerical example and an application to the $N$-machine power system with steam valve control.
Uncontrolled EV charging can significantly increase network peak demand, deteriorate bus voltage profiles, and increase operational costs in distribution networks. To mitigate these issues, this article proposes a multiagent – multilevel coordination based transactive EV charging framework in an active distribution network considering a three-level hierarchy of agents, i.e., distribution system operator, EV aggregators, and EV users. This article proposes a day-ahead optimal EV charging through coordinated transactive settlement and scheduling operations. In this work, the multiobjective optimization problem is formulated by providing autonomy to agents for considering the conflicting objectives such as charging cost minimization, revenue maximization, energy delivery maximization, and power mismatch minimization. The proposed framework is implemented on a modified IEEE 33 bus distribution network, a 118 bus distribution system, and an LV 907 bus European benchmark distribution network. The results show that the EV charging requirements are temporally distributed while maintaining network and EV users constraints and facilitating the objectives of agents. The proposed framework is compared with the state-of-the-art methods to showcase its techno-economic benefits, execution time, and convergence effectiveness. For the given test network, the charging cost reduction in the range of 24% to 55% is observed as compared to other methods used for comparative analysis, whereas the voltage profile is maintained above 0.9 pu. In addition, the performance of the proposed framework is measured using an economic utility metric.
Power system frequency is traditionally assumed to be spatially uniform, an assumption that underpins many analyses and operational frameworks. However, recent research has raised questions regarding the validity of this assumption under high penetration of inverter-based generators (IBGs). This article investigates spatial frequency heterogeneity in large-scale power systems using comprehensive time-domain simulations of a realistic synthetic model of the mainland Australian National Electricity Market (NEM). Established frequency heterogeneity analysis methods are applied to quantify both interarea and intraarea frequency variations following disturbances. The results show that spatial frequency variations in the NEM are nonnegligible, with their magnitude and spatial distribution strongly influenced by network topologies. Both interarea and intraarea frequency variations are observed, highlighting limitations of uniform frequency assumption for regional analyses. The proposed study further demonstrates that areas with high penetration of IBGs exhibit distinct spatial and temporal frequency heterogeneity characteristics compared to areas dominated by synchronous generators. These differences highlight the combined influence of network structure and generation technology on spatial frequency distribution. The research findings provide quantitative insight into the nature of frequency heterogeneity in large, radially connected power systems and highlight the limitations of uniform frequency assumption under increasing penetration of IBGs.
The rise of distributed renewable energy and the growing demand for electric vehicle (EV) charging create substantial operational unpredictability within residential energy systems designed to support the transition to electric transport. This work proposes a risk-aware, scenario-based multiobjective scheduling framework designed for the coordinated scheduling of residential microgrids integrating photovoltaic systems, battery storage systems, and shared EV charging infrastructure. The system is modeled as a unified energy-transportation framework that simultaneously balances household demand and EV flexibility while ensuring adherence to battery health requirements and utility grid constraints under environmental and demand uncertainties. A scenario-based extension of the Nondominated Sorting Genetic Algorithm-II was developed to compare deterministic scheduling, expected-value optimization, and a risk-aware formulation incorporating the conditional value-at-risk of an unmet critical load. The proposed method simultaneously minimizes the operating cost, battery degradation stress, grid dependency, reliability risk, and EV charging dissatisfaction over a 24-h horizon. Simulation results under representative sunny and nonsunny operating regimes demonstrate that the risk-aware framework reduces the mean unmet critical load by about 40% and the worst-case unmet load by nearly 44% under low-renewable conditions while achieving a 12.6% reduction in operating cost compared with deterministic scheduling. These results confirm that coordinated risk-aware scheduling significantly improves service reliability and transportation energy resilience under uncertainty while maintaining competitive economic performance.
A marking is a home state if it is reachable from every other reachable marking. Current research on home state analysis of unbounded Petri nets focuses solely on decision problems, while effective methods for home state exploration remain underdeveloped. This article proposes a home states analysis method of $\omega$-independent unbounded Petri nets. First, it presents the comprehensive method to explore all home states within the extended new modified reachability graph, which fully characterizes reachability between any markings. Second, it significantly enhances the efficiency of reachability analysis by avoiding the exploration of duplicate paths and applying negative circuits. Thirdly, the proposed method is formally supported by a theorem that establishes its validity. Finally, this article demonstrates the proposed method through a practical case.
In this article, two categories of output synchronization problems for coupled neural networks are tackled, that is, the cases with multiple output couplings and with multiple output derivative couplings. An output synchronization criterion for coupled neural networks with multiple output couplings is derived by employing the devised adaptive event-triggered control scheme, and Zeno behavior problem is also addressed. Moreover, the adaptive and event-triggered control methods are also utilized to deal with the output synchronization of coupled neural networks with multiple output derivative couplings, and the nonexistence of Zeno behavior is illustrated. Finally, two numerical examples, where the outputs of all nodes can realize the synchronization under the devised adaptive event-triggered strategies and Zeno behavior does not occur, are given to show the advantages of the proposed control schemes.
With large amounts of renewable energy, typical cyber attacks and nonideal operating conditions coexisting, the adaptive ability of the microgrid control based on physical models is insufficient due to the heavy reliance on specific optimization models, which need to be smart in different situations. This article proposes a memory-augmented reinforcement learning framework, in which the state-adaptive input weighting mechanism is innovatively developed to dynamically determine the significance of different state components, and then the recurrent neural network with long sequence processing capability is embedded to capture the long-term state dependencies, thereby augmenting memory and situation awareness capability. The secondary control problem is modeled as a Markov decision process, in which the knowledge-assisted reward design, integrated with staged training and adjusted exploration, jointly enhances the robustness and adaptability of microgrid secondary control. Simulation results of four cases show that the proposed control strategy significantly reduces the system frequency and voltage fluctuations. Compared with other strategies, it performs the best in terms of standard deviation, average deviation and excellent rate indicators.
This article proposes an integrated model-based and data-driven collaborative detection method for false data injection (FDI) attacks in multiarea power systems (MAPS). First, a model-based rapid detection layer is established using a finite-frequency $H_-/H_\infty$ observer. Its stability conditions are formulated as linear matrix inequalities via the generalized Kalman–Yakubovich–Popov lemma, which enhances attack sensitivity while maintaining robustness against disturbances. Second, a data-driven deep evaluation layer is developed based on conformal prediction with PID-based adaptive quantile adjustment to process the frequency-deviation data of each area. Prediction intervals and coverage variations are employed by the deep evaluation layer to support collaborative attack identification. The deep evaluation results are fed back to the rapid detection layer, where adaptive threshold adjustment improves FDI attack identification and reduces false alarms. Finally, the proposed method is validated on a three-area load frequency control system. Comparative studies with a recent data-driven two-stage detection method and a residual-threshold method demonstrate that the proposed method achieves a better balance between false alarms and missed detections while maintaining high detection sensitivity and accuracy under multifrequency and multiarea FDI attacks.
