Modular methods, which decompose complex joint policies into function-specific sub-policies, have been widely adopted to enhance asymptotic performance in single-task cooperative multi-agent reinforcement learning (MARL). However, modular policies trained on source tasks often struggle to generalize to unseen scenarios due to variations across tasks, such as mismatched action spaces and divergent state dynamics. To address this challenge, we propose Mutual Information-Guided Subtask Selection(MIGSS), a novel framework that enhances zero-shot generalization in MARL through two key innovations: a Discriminative Group Trajectory Encoder and Global Attention-Driven Coordination. Specifically, the Discriminative Group Trajectory Encoder remaps agent trajectories by maximizing mutual information between agent trajectories and dynamically assigned groups. This optimizes cross-task consistent group trajectory with broader embedding distributions. This encourages agents in distinct states to select specialized subtasks, effectively promoting functional modularity. Meanwhile, the Global Attention-Driven Coordination employs a global attention mechanism to integrate state information, coordinating group trajectories for expressive credit assignment. Extensive experiments in StarCraft II cooperative scenarios demonstrate that MIGSS significantly outperforms superior zero-shot generalization baselines in both single-task and multi-task settings.Visualization analyses confirm that the learned group trajectories successfully disperse agent trajectories into a consistent and broader embedding space, thereby enhancing subtask modularization.

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.

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.

Unmanned Aerial Vehicles (UAVs) are emerging as a pivotal component in the field of intelligent transportation systems. Leveraging virtual-physical interactions, digital twin technology significantly enhances the adaptability of UAVs in complex traffic environments. However, current approaches still pose three major challenges: contextual adaptability, timely responsiveness, and effective multi-UAV coordination. In this paper, we introduce EnFlexiTwin, a digital twin enhancement assistance platform seamlessly integrated with AdaSor, a lightweight adaptive data selector. EnFlexiTwin automates the construction of incremental learning datasets, enabling real-time enhancements that allow digital twins to adapt to new time series data while preserving historical knowledge. We test EnFlexiTwin on a real-world dataset from low-altitude small-parcel delivery. The results show improved performance and adaptability of digital twins. Furthermore, time-varying simulations on real-world dataset and experiments on a practical air-ground cooperative UAV swarm application highlight that EnFlexiTwin achieves superior enhancements under varying real-time requirements and swarm scale compared to baseline approaches.

The Industrial Internet of Things (IIoT) has become a critical technology to accelerate the process of digital and intelligent transformation of industries. As the cooperative relationship between smart devices in IIoT becomes more complex, obtaining deterministic responses of IIoT periodic time-critical computing tasks becomes a crucial and nontrivial problem. However, few current works in cloud/edge/fog computing focus on this problem. This paper is a pioneer in exploring deterministic scheduling and network structural optimization problems for IIoT periodic time-critical computing tasks. We first formulate the two problems and derive theorems to help quickly identify computation and network resource sharing conflicts. Based on this, we propose a deterministic scheduling algorithm,IIoTBroker, which realizes a deterministic response for each IIoT task by optimizing the fine-grained computation and network resources, and a network optimization algorithm,IIoTDeployer, which provides a cost-effective structural upgrade solution for existing IIoT networks. Our methods are illustrated to be cost-friendly, scalable, and deterministic response guaranteed with low computation cost from our simulation results.

A neural processing unit (NPU) is a microprocessor which is specially designed for various types of neural network applications. Because of its high acceleration efficiency and lower power consumption, the airborne embedded system has widely deployed NPU to replace GPU as the new accelerator. Unfortunately, the inherent scheduler of NPU does not consider real-time scheduling. Therefore, it cannot meet real-time requirements of airborne embedded systems. At present, there is less research on the multi-task real-time scheduling of the NPU device. In this article, we first design an NPU resource management framework based on Kubernetes. Then, we propose WAMSPRES, a workload-aware NPU performance model based soft preemptive real-time scheduling method. The proposed workload-aware NPU performance model can accurately predict the remaining execution time of the task when it runs with other tasks concurrently. The soft preemptive real-time scheduling algorithm can provide approximate preemption capability by dynamically adjusting the NPU computing resources of tasks. Finally, we implement a prototype NPU scheduler of the airborne embedded system for the fixed-wing UAV. The proposed models and algorithms are validated on both the simulated and realistic task sets. Experimental results illustrate that WAMSPRES can achieve low prediction error and high scheduling success rate.

Visual object tracking is a very important task for unmanned aerial vehicle (UAV). Limited resources of UAV lead to strong demand for efficient and robust trackers. In recent years, deep learning-based trackers, especially, siamese trackers achieve very impressive results. Though siamese trackers can run a relatively fast speed on the high-end GPU, they are becoming heavier and heavier which restricts them to be deployed on UAV platform. In this work, we propose a lightweight aerial tracker based on the siamese network. We use EfficientNet as the backbone, which has less parameters and stronger feature extract ability compared with ResNet-50. After a pixel-wise correlation, a classification branch and a regression branch are applied to predict the front/back score and offset of the target without the predefined anchor. The results show that our tracker works efficiently and achieves impressive performance on UAV tracking datasets. In addition, the real-world test shows that it runs effectively on the Nvidia Jetson NX deployed on DJI UAV.

This work aims to explore the effect of Resource Allocation and Control (RAC) of the Unmanned Aerial Vehicle (UAV)-assisted Internet of Things (IoT) communication equipment. The communication congestion problem is caused by insufficient computing resources of macro base stations when many user terminals access the IoT system. To this end, this work innovatively constructs the IoT system combined with UAVs to satisfy the user’s communication experience and allocate channel resources simultaneously. Finally, a UAV-assisted IoT resource allocation and control model is constructed and tested through experiments on the simulation software to analyze its performance. The results indicate that the transmission delay of the model reported here is significantly lower than that of other algorithms with the increased number of end users and the proportion of tasks. When the number of users is 500, the transmission delay of data resources is 47.5s. Besides, it requires lower energy consumption and shorter task completion time while maintaining lower throughput. At the same time, the uplink transmission rate of the model is generally stable at 3.10bps; the uplink transmission rate of the entire wireless communication system shows an increasing trend with the increase of the transmit power. Therefore, it is found that the model constructed here has strong robustness and achieves an excellent data resource allocation effect while ensuring lower data delay. This model provides an experimental reference and contribution to the improvement of the performance of IoT systems and the accurate allocation of network resources.

To further improve the capacity of airborne embedded system for dealing with deep learning (DL) applications and reduce overall power consumption, it is necessary to equip Neural Processing Units (NPUs). Comparing with the cloud system, the airborne embedded system usually has a fixed application set, but strict real-time constraints. Unfortunately, the inherent NPU scheduler does not consider the application priority, which cannot provide the sufficient real-time capability for the airborne embedded system. At present, there are few researches on multi-task real-time scheduling for NPUs. Therefore, we propose WMDRS, a workload-aware performance model multi-task dynamic-quota real-time scheduling for Neural Processing Units. The NPU performance model based on workload-awareness can accurately predict the remaining execution time of a task, which is running concurrently with other tasks on NPU. The multi-task dynamic-quota real-time scheduling algorithm can provide the approximate preemption by dynamically adjusting NPU computing resources for active applications. In addition, we implement a prototype NPU scheduler without any hardware extension. Furthermore, the proposed NPU performance model and real-time scheduling algorithm are evaluated in realistic application sets. Experimental results demonstrate that WMDRS can achieve low prediction error and high scheduling success ratio.

New generation airborne embedded system has deployed Graphical Processing Units (GPUs) to raise processing capability to meet growing computational demands. Comparing with the cloud system, the airborne embedded system usually has a fixed application set, but strict real-time constraints. Unfortunately, the inherent GPU scheduler does not consider the application priority, which cannot provide the sufficient real-time capability to the airborne embedded system. To meet timeliness requirements, it is necessary to predict timing behaviors of those applications and design a real-time scheduling policy based on priority and deadline. We therefore propose WAMP2S, a workload-aware GPU performance model based pseudo-preemptive real-time scheduling algorithm for the airborne embedded system. The workload-aware GPU performance model can accurately predict the execution time of an application, which is running concurrently with other applications on GPU. The pseudo-preemptive real-time scheduling algorithm can provide the approximate preemption by dynamically adjusting GPU computing resources for active applications. Unlike previous work on GPU performance model and GPU real-time scheduling, WAMP2S considers the impact of co-executing workload on the execution time estimation and provides a software-only approach for preemption support. In addition, WAMP2S implements a prototype GPU scheduler without any source code analysis. We evaluate the proposed GPU performance model and real-time scheduling algorithm in both simulated and realistic application sets. Experimental results illustrate that WAMP2S can achieve low prediction error and high scheduling success ratio.

Intelligent unmanned systems (IUSs) are distributed systems composed of multiple agents that share information or cooperate to accomplish specific complex tasks. Agents of the IUS are capable of perception, cognition, control, decision-making, and action. In some cases, the environmental situation and task objectives faced by the IUSs are constantly changing with time. Thus, IUSs are time-sensitive systems. To accelerate the task execution time and response speed, IUSs use artificial intelligence technology to increase the speed and quality of the `observation-orientation-decision-action' (OODA) cycle of task execution. IUSs will tend to decompose the system into different functional units in the future, and individuals take different task roles from the functional perspective of OODA. The system is evolving from a linear OODA cycle of individuals to a cooperative OODA (Co-OODA) with different node roles. At present, the reinforcement learning (RL) algorithm is the mainstream method to solve IUSs cooperation problems. However, it does not adapt to the Co-OODA with different roles; and cannot maximize the Co-OODA system's potential. This paper introduces the role-based Co-OODA system. Furthermore, we propose and design a role-based deep reinforcement learning framework and its corresponding information sharing mechanism.

New generation airborne embedded system has deployed Graphical Processing Units (GPUs) to raise processing capability to meet growing computational demands. Applications in the airborne embedded system have strict real-time constraints. Therefore, it is necessary to accurately predict timing behaviors of those applications. Many previous work propose GPU performance models to estimate the execution time of applications. However, most of those models do not consider the impact of co-execution on the GPU performance. In this paper, we propose a workload-aware GPU performance model to predict the execution time of applications executed concurrently on a single GPU. Experimental results illustrate that the proposed model can achieve a 5.1%-11.6% prediction error in a real airborne embedded hardware platform.

Vehicularad hocnetworks (VANETs) have far-reaching application potentials in the intelligent transportation system (ITS) such as traffic management, accident avoidance and in-car infotainment. However, security has always been a challenge to VANETs, which may cause severe harm to the ITS. Sybil attack is considered as a serious security threat to VANETs since the adversary can disseminate false messages with multiple forged identities to attack various applications in the ITS. RSSI-based Sybil nodes detection is an efficient scheme against Sybil attacks, which adopts position estimation, distribution verification or similarity comparison to identify Sybil nodes. However, when Sybil nodes conduct power control to deliberately change transmission powers, the received RSSI values would change correspondingly, which leads to inaccurate localization or different RSSI time series of these Sybil nodes. Thus, it is very difficult to differentiate Sybil nodes from normal nodes via conventional RSSI-based methods. This paper first discusses potential power control models (PCMs) for launching Sybil attacks in VANETs, then presents two simple Sybil attack models and three sophisticated Sybil attack ones with or without power control in detail, finally proposes a power control identification Sybil attack detection (PCISAD) scheme to find anomalous variations in RSSI time series, which are then used to identify Sybil nodes via a linear SVM classifier. Extensive simulations and real-world experiments prove that the proposed scheme can effectively deal with Sybil attacks with power control.

Cyber physical systems (CPSs) incorporate computation, communication, and physical processes. The deep coupling and continuous interaction between such processes lead to a significant increase in complexity in the design and implementation of CPSs. Consequently, whereas developing CPSs from scratch is inefficient, developing them with the aid of CPS run-time sup-porting platforms can be efficient. In recent years, much research has been actively conducted on CPS run-time supporting plat-forms. However, few surveys have been conducted on these platforms. In this paper, we analyze and evaluate existing CPS run-time supporting platforms by first classifying them into three categories from the viewpoint of software architecture: com-ponent-based platforms, service-based platforms, and agent-based platforms. Then, for each type, we detail its design philosophy, key technical problems, and corresponding solutions with specific use cases. Subsequently, we compare existing platforms from two aspects: construction approaches for CPS tasks and support for non-functional properties. Finally, we outline several im-portant future research issues.

Cyber-Physical Systems (CPSs) are interactive, intelligent and distributed-hybrid systems which have computing units embedded in physical environment and widely applied in the safety-critical field. Compared with the traditional embedded hybrid system, the problems of safety, reliability and uncertainty, caused by constant interaction between computing and physical process, are more prominent than ever before. An astronautic case has been taken for example in this paper. Correspondingly, the Discrete Hybrid Automata (DHA) modeling frame and Hybrid System Description Language (HYSDEL) are adopted to build and analyze its behavior model. Besides, combined with the hybrid toolbox, the trajectories of the continuous states and the reachability of system are simulated and analyzed. The usage of the approach to modeling and analysis of CPS has been applied in the scene of lunar rover autonomous walking, which lay a model foundation for the further safety verification.

Resource performance monitoring is among the most active research topics in distributed computing. In this paper, we propose an adaptive resource monitoring method for applications in heterogeneous computing environment. According to the operating environment of distributed heterogeneous system and the changes of system resource workload, the method combines periodic pull mode with event-driven push mode to adaptively publish and retrieve system resource information. Preliminary experiments reveal that, by using our adaptive monitoring method, the efficiency of system monitoring is improved over that accrued by using regular monitoring approaches.