Unmanned swarms have widely penetrated and profoundly affected all fields of modern society. In the process of mission execution, unmanned swarms may face a range of interferences and damages, potentially affecting their normal operation and mission efficiency. Therefore, evaluating the robustness of swarm has important theoretical and applied value to ensure the stability and efficiency of the swarm in the dynamic environment. This paper focuses on cooperative sensing of the UAV swarm as an example. Based on behavioral characteristics of swarm cooperation, the cooperation process is modeled using the Hawk and Dove evolutionary game. A robustness indicator set of UAV swarm is constructed from the three dimensions: swarm attributions, environmental attributions, and task effects. Quantitative robustness metrics are proposed, and the robustness performance of the swarm under various perturbation factors is analyzed. An intelligent evaluation model eXGBoost based on the Shapley additive explanation and the eXtreme Gradient Boosting model is proposed. The experimental results show that the proposed robustness evaluation method is both feasible and effective. In addition, the method increases the transparency of the evaluation model, providing effective feedback for the design of unmanned swarms.

Automated Program Repair (APR) aims to resolve software bugs without human intervention, but complex logic errors and silent failures remain challenging. Existing LLM-based APR methods mainly rely on source code and coarse test feedback, making it difficult to capture runtime behaviors and dynamic data dependencies. Execution traces expose concrete state transitions, yet a single LLM interpreting them in isolation may commit to an incorrect repair hypothesis and produce test-overfitting patches. We therefore treat runtime evidence as shared constraints for validating repair hypotheses rather than merely as additional prompt context. We propose TraceRepair, a multi-agent framework in which a Probe Agent captures execution snapshots of critical variables, while specialized repair agents generate, compare, and iteratively refine candidate patches against the observed runtime evidence. A Judge Agent then arbitrates the remaining hypotheses and synthesizes the final patch. Evaluated on Defects4J, TraceRepair correctly fixes 392 defects and outperforms existing LLM-based approaches. Further experiments demonstrate improved efficiency and strong generalization on a newly constructed dataset of recent bugs, suggesting that the gains arise from dynamic reasoning rather than memorization.

Context: Automated Program Repair (APR) based on Large Language Models (LLMs) has made rapid progress in recent years. Existing methods are no longer limited to direct patch generation; many of them introduce reasoning traces, feedback, or contextual information before repair. However, such information is often collected in a coarse-grained and weakly targeted manner, making it difficult for the model to distinguish domain constraints, intended program behavior, and the root cause of the defect. As a result, their repair performance remains limited when handling complex bugs. Objective: Human programmers usually do not fix bugs by directly editing code. Instead, they often follow a diagnostic process: understanding the relevant domain context, clarifying the intended functionality of the code, and then analyzing why the current implementation violates the expected behavior. This process fits the nature of program repair, because a correct patch should satisfy domain constraints, preserve intended functionality, and address the root cause of the failure. Inspired by this observation, we propose CAMP, a Cognition-Aligned Multi-stage APR framework, which organizes pre-repair reasoning according to this diagnostic process. We also investigate whether such structured reasoning can generalize to different programming languages and recent real-world bugs. Methods: CAMP consists of two stages: Chain-of-Thought (CoT) collection and bug repair. In the CoT collection stage, CAMP explicitly decomposes pre-repair reasoning into three progressive steps: domain knowledge extraction, code functionality analysis, and bug cause analysis. The refined reasoning information is then integrated into the repair prompt to guide patch generation. We compare CAMP with 10 state-of-the-art APR methods on Defects4J V1.2 and V2.0. To evaluate generalization and practical applicability, we further test CAMP on DebugBench, a multilingual dataset covering C++, Java, and Python, and on RecentAPR-98, a newly constructed dataset of recent real-world Java bugs. Results: CAMP repairs 120 bugs on Defects4J V1.2 and 109 bugs on Defects4J V2.0, achieving performance comparable to or better than existing state-of-the-art baselines. On DebugBench, CAMP shows strong repair capability across C++, Java, and Python, indicating good cross-language generalization. On RecentAPR-98, under the same experimental setting, CAMP repairs 8 more bugs than the previous state-of-the-art method. In terms of cost, CAMP consumes about 53,615 tokens per bug on average, corresponding to approximately $0.0095, reducing the cost by at least 93.2% compared with mainstream LLM-based repair methods. Conclusion: These results show that CAMP’s effectiveness does not come from simply increasing the number of reasoning steps, but from organizing repair-relevant information before patch generation. By separating domain knowledge, functionality understanding, and bug cause analysis, CAMP provides the model with more targeted repair evidence, thereby improving repair effectiveness, generalization, and cost efficiency.

Artificial Intelligence of Things (AIoT) aims to build a self-learning, self-adaptive, and self-evolving Internet of Things ecosystem, which has facilitated many promising intelligent services. Data is an important foundational element for many applications. Establishing a well-designed trading mechanism to collect the necessary data from various sources is essential to realize the vision of AIoT. In this article we investigate the data trading incentive mechanism between multiproviders and multibuyers for AIoT. To address the two-sided dilemma, we develop a joint optimization game to maximize the payoff of all market participants. A two-layer Stackelberg evolutionary game theoretic framework is developed to divide the optimization problem into two subproblems: one for data pricing by providers and the other for purchasing decisions by buyers. The subproblem of optimal data pricing for providers is modeled as a noncooperative game. Providers utilize the game's equilibrium solution to dynamically modify their pricing strategies in response to a changing competitive environment and demanding. This is because buyers have limited information, their behaviors are modeled via evolutionary game. By encouraging data providers to take the buyers' evolutionary dynamics into account has the potential to overcome the myopia behaviors. The equilibrium solution is obtained via replicator dynamics. Extensive experiments demonstrate the efficacy and efficiency of the proposed hierarchical interaction framework. Overall, our results show the proposed Stackelberg evolutionary game framework establishes a desired data market and achieves higher long-term revenue for both sides of participants in the market. The hierarchical framework can effectively prompts data trading in the market.

With the increasing scale and complexity of power systems, the Security and Stability Control System (SSCS) plays a vital role in ensuring the safe operation of the grid. However, existing SSCS implementations still face many limitations in cross-regional coordination, control precision, and risk prediction. Establishing the digital simulation model is an effective way to verify the control policy of SSCS. This paper proposes a neural heuristic task scheduling method based on deep reinforcement learning (DRL) to schedule the simulation tasks. It models the task dependencies of SSCS as a directed acyclic graph (DAG) and then dynamically optimizes task priorities and resource allocation through deep reinforcement learning. The method introduces multi-head attention and heterogeneous attention mechanisms to effectively capture complex dependencies among tasks, enabling efficient multi-core task scheduling. Simulation results show that the proposed algorithm significantly outperforms traditional scheduling methods in terms of makespan, load balancing, and resource utilization. It can also adapt to dynamic changes under different task scales and multi-core environments, demonstrating strong robustness and scalability.

Multi-access edge computing is an effective computing paradigm that supports computation-intensive applications for Internet of Things (IoT) equipment. Due to the shortage of environmental knowledge, it is difficult for each device to decide the optimal offloading decision in a distributed way. The dynamic nature of task generation and the diversity of requirements also present a significant challenge to the resource management of edge servers. The problem is additionally complex, as the two decisions are coupled to determine the utility of IoT devices. To overcome these challenges, the paper addresses the online computation offloading and dynamic task scheduling issue in multi-access IoT. We establish a two-layer adaptive approach called DENOTE to address the problem. Benefiting from evolutionary game and deep reinforcement learning, we propose a service provider selection game based on replicator dynamics to make offloading decisions at the device side. We also present a proximal policy optimization based approach to schedule the heterogeneous tasks at the edge side. Extensive experiments are carried out to assess the proposed approach in terms of the convergence speed of DENOTE and long-term cost of devices. The findings indicate the approach has superior convergence and outperforms baselines with less information.

Unmanned Aerial Vehicles (UAVs) are playing an increasingly critical role in Internet of Things (IoT), and with AI advancements, they are evolving into intelligent, multi-functional platforms supporting crowdsensing applications in modern IoT ecosystems. Given the dynamic nature of UAV-based IoT services, it is crucial to evaluate whether UAVs meet essential performance characteristics like flexibility, robustness, and stability. Flight data of UAVs is fundamental for assessing these characteristics, and effective data management is key to enabling timely, accurate performance evaluations. In this paper, we propose a Hierarchical Tree Model (HTM) specifically designed to accommodate the characteristics of UAVs-generated data. This model is supported by a hierarchical tree-based storage structure that optimizes the organization and retrieval of time-series data. To enable characteristic evaluation, we use tags to store characteristic information. To enable characteristic evaluation, we use tags to store characteristic information and apply encoding/decoding algorithms for flexible operation. We extend the SQL syntax tree of IoTDB with new syntax and semantics, enhancing the IoTDB parser and optimizer for tag integration and data operation. Experimental results show that our framework enables more efficient real-time evaluation of UAVs performance and better scalability for large deployments. Our findings highlight the potential of this time-series data management approach to support the real-time evaluation of UAVs characteristics, facilitating more informed decision-making and resource allocation in IoT-driven environments.

The increasing integration of renewable energy sources, such as wind and solar, into power grids has amplified the demand for effective management of Security and Stability Control System (SSCS). To address cost and scalability challenges, digital simulations have become indispensable for evaluating the control strategies. As grid complexity grows, the need for precise simulations and efficient task scheduling in multi-core environments has become important. This paper introduces a memory-priority scheduling algorithm for the simulations of SSCS model, designed to optimize memory access and enhance task handling efficiency. We also derive the upper bound of the response time. Comparative analysis reveal its superior performance over existing algorithms, particularly in dynamic power system environments.

Uncrewed aerial vehicles (UAVs) are playing an increasingly critical role in Internet of Things (IoT), and with AI advancements, they are evolving into intelligent, multifunctional platforms supporting crowdsensing applications in modern IoT ecosystems. Given the dynamic nature of UAV-based IoT services, it is crucial to evaluate whether UAVs meet essential performance characteristics like flexibility, robustness, and stability. Flight data of UAVs is fundamental for assessing these characteristics, and effective data management is key to enabling timely, accurate performance evaluations. In this article, we propose a hierarchical tree model (HTM) specifically designed to accommodate the characteristics of UAVs-generated data. This model is supported by a hierarchical tree-based storage structure that optimizes the organization and retrieval of time-series data. To enable characteristic evaluation, we use tags to store characteristic information. To enable characteristic evaluation, we use tags to store characteristic information and apply encoding/decoding algorithms for flexible operation. We extend the SQL syntax tree of IoTDB with new syntax and semantics, enhancing the IoTDB parser and optimizer for tag integration and data operation. Experimental results show that our framework enables more efficient real-time evaluation of UAVs performance and better scalability for large deployments. Our findings highlight the potential of this time-series data management approach to support the real-time evaluation of UAVs characteristics, facilitating more informed decision-making and resource allocation in IoT-driven environments.

In recent years, the large-scale integration of new energy sources, such as wind and photovoltaic power into power grid, has necessitated higher standards for the safe, stable, and flexible control of power systems. The Security and Stability Control System (SSCS) is proposed to monitor the operation of the grid. To address the cost and scale limitations, simulation has been emerging as a crucial method for analyzing safety and stability control systems. However, as the grid’s complexity increases, the precision and computation workload of the SSCS associated to the grid simulation model have significantly grown. Real-time task scheduling in multi-core systems is important for detecting abnormal events of complex stability control simulation models in time. This paper introduces a new scheduling algorithm specifically designed for the simulation model in power system. It also proposes an interference analysis method to address delays caused by contention for shared hardware resources. We conduct extensive experiments to analyze the proposed algorithm. Compared with existing scheduling algorithms, the scheduling algorithm of the paper achieves superior performance and effectiveness.

The Security and Stability Control System (SSCS) is developed to ensure the safe and stable operation of power grids, effectively mitigating failure propagation through emergency control strategies. However, due to the diversity, complexity, and region-specific nature of SSCS, the need for individualized development often arises, leading to time-consuming and error-prone in SSCS development. To address this challenge, we introduce ERC-Code, an automated development framework that empowers developers to prioritize design over implementation. ERC-Code includes extended reactive component modeling, correctness specifications, and model-to-code transformations. The extended reactive component offers the advantage of capturing the complex logic of SSCS at a high level of abstraction, including loops, branches and synchronizations. The correctness specifications prevent the creation of incomplete and invalid models, ensuring the robustness of the generated code. Additionally, the model-to-code engine automatically generates code from the extended reactive component. Our evaluation covers three key aspects: firstly, we assessed the effectiveness of ERC-Code in three real-world SSCS applications. We then conducted a comparison between the traces generated by ERC-Code and the ERC model. Finally, we conducted an experiment to evaluate developer productivity, drawing comparisons between ERC-Code and a code-centric approach. Our results exhibit promise, affirming ERC-Code’s applicability and code quality, along with enhanced productivity.

The change of illumination has a great impact on the imaging effect of visible image. In the tracking scene of bad illumination or weather conditions, visible image cannot effectively and comprehensively describe the imaging scene. It is important to consider the effect of illumination change in the process of feature extraction and fusion, and fuse complementary information from visible modality and thermal infrared modality to provide a comprehensive description of the scene. Aiming to enhance the performance of nighttime object detection and tracking, this paper designs an illumination-aware module, which utilizes the illumination information contained in the visible image. The module generates the fusion weights for the visible and thermal infrared images. Then the multi-modal fusion algorithm based on the module for object tracking is proposed. Extensive experiments have shown that the proposed illumination-aware multi-modal fusion effectively improves the performance of object tracking in the nighttime environment.

A huge amount of multimedia data is generated everyday. The value of data determines that it can be viewed as a kind of commodity for sale. Processing raw data to extract knowledge can increase the value of data. However, processing cost determine the profits of data providers. Processing delay affect the purchasing willingness of data buyers. In this paper, we propose an economics framework for data trading. A data market with the aid of edge computing is established to describe the interaction between data providers and data buyers. The problem of pricing and purchasing data is discussed to maximize the profits of the whole participants in multimedia data market. A Stackelberg game is then proposed to derive the best pricing and purchasing strategies. The knapsack strategy is also used to select the candidate data providers. Extensive experimental results show the effectiveness of the proposed framework.

Unmanned swarm systems are now employed in a wide range of applications. Agriculture, commerce, and the military will all benefit from the advancement of unmanned swarm technology. Scenarios should have a standardized description and definition as system input to support numerous applications and accelerate scenario validation, allowing for more precise and specialized processing to standardize the validation process. A scenario description language for unmanned swarm (SL4U) is introduced in order to implement the unmanned swarm scenario. The scenario is divided into environment and task. The environment is defined, including the task space, target, and home locations, and the initial settings of the unmanned swarm. In terms of tasks, four categories of tasks, four temporal relationship constraints, the event handler, as well as other elements are identified. The dual engines of parsing and micro-embedded collaboration we developed will handle the SL4U. We employ a search and rescue scenario in both simulation and physics to test SL4U. Through the processing of the dual engines, the unmanned swarm’s search and rescue operations in a predefined environment are accomplished successfully. In more complex scenarios, just add more descriptions to SL4U and implement them in dual engines rather than starting from scratch. The use of SL4U in verification minimizes the cost and time consumption of unmanned swarm, particularly in physical verification.

Recent advances in artificial intelligence (AI) have impacts on the development of communication networks. Diversified applications and scenarios put forward AI-native requirement for communication network. The network is envisioned to offer high data rate and low-latency communications, pervasive artificial intelligence, full coverage and programmable service. With the aid of AI, autonomous unmanned aerial vehicles (UAVs) play an indispensable role in the emerging scenarios. Owing to their flexibility and scalability, UAVs can be utilized as the mobile platform to deliver various kinds of services in dynamic and adversarial environments, especially the emergency scenarios. In this article, we deal with the post-disaster scenarios that UAVs, equipped with computing servers, work with the remaining infrastructures to provide communication and computing services for the ground mobile users (GMUs). With the aim of minimizing the energy cost and computing delay of task offloading, a two-layer optimization mechanism that can dynamically orchestrate the task offloading decision and the UAVs deployment is proposed. In lower layer, the problem of making offloading decisions is formulated into an evolutionary game. Replicator dynamics is exploited to help users to make decisions. In upper layer, a clustering approach for users is introduced to avoid unbalanced load distribution. By suggesting the locations and number of UAVs based the clusters, system cost is further reduced. The fairness of service load among UAVs is also achieved. Numerical results demonstrate the effectiveness of the proposed two-layer optimization mechanism.

UAVs need to use a large number of sensors to complete autonomous navigation tasks in indoor environments. The evaluation of multi-sensors ranging capabilities in different application scenarios is an important research direction at present. This paper analyzes three main factors (light intensity, distance, target material) affecting the ranging capability of three sensors (depth camera D435i, binocular camera T265, and lidar Rplidar A1). To quantify the ranging capability of sensors in different application scenarios, the concept of dynamic ranging capability index (DRCI) is proposed in this study, which mainly depends on ranging accuracy. DRCI is a fuzzy evaluation index. The higher DRCI value, the stronger ranging capability. Through the analysis of a large number of experimental data, we establish the DRCI calculation method. This paper uses three sensors for a ranging task while calculating DRCI by the method in Airsim simulation scenarios and real scenarios. Finally, the real-time changes in the ranging capability of multi-sensors can be effectively evaluated by the method, and the results in simulation scenarios are validated in real scenarios.

随着无人系统、智能技术的发展以及应用日趋复杂,由若干无人系统在一定时间与空间内自主协同完成任务的智能无人集群系统得以快速涌现,其可有效提高完成复杂任务的能力,在国防装备与国民经济等领域具有重要且广泛的应用前景.无人集群系统的智能评价具有场景多样、指标多维、方法多类等特征,是极具挑战性的研究问题.本文在分析无人集群系统智能评价已有相关研究基础上,结合国家科研项目,基于群体观察-判断-决策-执行(OODA)行为模型提出了评价协同演化模型及适应群体OODA的等级评价标准,进而提出场景驱动的评价过程与适应性评价方法,设计开发自主评价工具集,并结合无人集群系统(USS)及其典型应用场景进行了初步验证.

This paper presents a method to verify the safety of CPS based on theorem proving. First, the formal reasoning mechanism of theorem proving technology is discussed, and then the theorem proving tool KeY maera and its operation model Hybrid Program are introduced. Finally, the temperature control system is modeled by HP and input the model programs into KeYmaera. Based on the automatic reasoning function of the tool, the safety of temperature control is verified. The correctness and effectiveness of the CPS safety verification method are illustrated.

Swarm robots are composed of a certain number of physical robots. Through interaction and cooperation, individuals in this kind of system can sprout robust, scalable and flexible swarm intelligence behavior. Behavior modeling of swarm robots is one of the basic methods for scientific analysis of such systems, and it is also an important basis for its capability evaluation. The research and analysis of behavior model of swarm robots is helpful to understand its behavior mechanism. Firstly, this paper summarizes the main characteristics of swarm robots, and lists and summarizes several typical spatial cooperation behaviors of swarm robots. On this basis, the spatial cooperation behavior of swarm robots is classified, and the modeling methods of spatial cooperation behavior of cluster robots are summarized and expounded, including graph theory, bionic model, dynamic model and learning model. Various modeling methods are analyzed and compared from the multi-dimensional aspects of method overview, application examples and applicable scenes. Finally, the problems of spatial cooperative behavior modeling of swarm robots to be further studied are put forward, and possible research directions in the future are prospected, so as to better support the evaluation and optimization design of swarm robots. The purpose of this paper is to enable relevant researchers to comprehensively and systematically understand the spatial cooperative behavior modeling method of swarm robots.

针对有人机/无人机编队中出现的任务负载不均衡、实时性不足等问题,结合任务协同特点及实际作战需求,对传统合同网协议进行扩展,提出了一种有人机/无人机任务协同方法.根据战场实际需求分析有人机/无人机任务协同过程,建立有人机/无人机任务协同模型,通过求解约束条件下的效能函数,确定效能参数,确保有人机/无人机任务编队达到最短执行时间及效能最大化.实验结果说明,基于扩展合同网协议的有人机/无人机任务协同方法能有效均衡有人机/无人机任务负载,减少任务执行时间,具有良好的实时性和扩展性.