ABSTRACT The rise in malware attacks on Internet of Things (IoT)‐based Android devices has significantly increased cybersecurity risks. To combat these threats, researchers have introduced machine learning (ML) and deep learning (DL) methods. However, the rapid spread of Android devices across diverse geographic areas has resulted in distributed data, rendering traditional methods suboptimal. Centralizing data not only raises privacy concerns but also introduces overhead and scalability issues. To address these challenges, the research community has developed federated learning (FL)‐based systems for Android malware classification, aiming to balance privacy and scalability without compromising effectiveness. In the standard FL paradigm, federated averaging (FedAvg) converges the local models of all participating clients into a global model in each iteration. A notable limitation of FedAvg, however, is that it does not differentiate between the contributions of individual local models, leading to potential performance degradation if suboptimal local models are included. This research introduces a dynamic weighted federated averaging (DW‐FedAvg) mechanism to overcome these limitations. DW‐FedAvg adjusts the weights of each local model based on its performance at the client level. The efficacy of DW‐FedAvg was evaluated using four benchmark datasets in Android malware classification. Preliminary results demonstrate that the proposed methodology outperforms traditional FedAvg and other state‐of‐the‐art techniques in terms of scalability, privacy preservation, and performance metrics such as accuracy, F1 score, AUC score, and FPR score.

Polarization images facilitate image enhancement and 3D reconstruction tasks, but the limited accessibility of polarization cameras hinders their broader application. This gap drives the need for synthesizing photorealistic polarization images. The existing polarization simulator Mitsuba relies on a parametric polarization image formation model and requires extensive 3D assets covering shape and PBR materials, preventing it from generating large-scale photorealistic images. To address this problem, we propose PolarAnything, capable of synthesizing polarization images from a single RGB input with both photorealism and physical accuracy, eliminating the dependency on 3D asset collections. Drawing inspiration from the zero-shot performance of pretrained diffusion models, we introduce a diffusion-based generative framework with an effective representation strategy that preserves the fidelity of polarization properties. Experiments show that our model generates high-quality polarization images and supports downstream tasks like shape from polarization.

Shape from polarization (SfP) benefits from advancements like polarization cameras for single-shot normal estimation, but its performance heavily relies on light conditions. This paper proposes SfPUEL, an end-to-end SfP method to jointly estimate surface normal and material under unknown environment light. To handle this challenging light condition, we design a transformer-based framework for enhancing the perception of global context features. We further propose to integrate photometric stereo (PS) priors from pretrained models to enrich extracted features for high-quality normal predictions. As metallic and dielectric materials exhibit different BRDFs, SfPUEL additionally predicts dielectric and metallic material segmentation to further boost performance. Experimental results on synthetic and our collected real-world dataset demonstrate that SfPUEL significantly outperforms existing SfP and single-shot normal estimation methods. The code and dataset is available at https://github.com/YouweiLyu/SfPUEL.

On the basis of analyzing the architecture and characteristics of Combat cloud and Tactical cloud, one resource-visualization-based coordinative computing architecture for Tactical cloud is focused on and studied. Further, several vital mechanisms are also studied and discussed, covering the servitization encapsulation of heterogeneous resources, management of service pool, mission passing and the dynamic mapping of service and subtasks. All these proposed architecture and mechanisms can be a good foundation for the constructing and realizing of Tactical clouds.

With the rapid increase in information quantity in the Internet of Things (IoT) era, traditional methods of acquiring information can no longer meet the present situation of diversified demand for information acquisition. In this paper, an open general intelligent IoT information service system based on Bluetooth Low Energy (BLE) tags, mobile devices and cloud services is studied and designed. On the one hand, the system deploys the BLE tag in the application object. On the other hand, it automatically selects the BLE tag in real time and obtains the corresponding information of the tag in the cloud through the self-developed Mini Program. Meanwhile, a data storage servicer is designed and developed based on the public cloud to store the BLE tag and the data of application object. After the final multi-scenario test, the system has good scalability, and has the characteristics of real-time, efficient and intelligent.

Vehicular Ad Hoc Networks (VANETs) bring many benefits and conveniences to road safety and drive comfort in future transportation systems. However, VANETs suffer from almost all security issues as same as wireless networks. Sybil attack is one of the most risky threats since it violates the fundamental assumption of VANETs-based applications that all received information are correct and trusted. Sybil attacker can generate multiple fake identities to disseminate false messages. In this paper, we propose a novel Sybil attack detection method based on Received Signal Strength Indicator (RSSI), Voiceprint, to conduct a widely applicable, lightweight and full-distributed detection for VANETs. Unlike most of previous RSSI-based methods that compute the absolute position or relative distance according to RSSI values, or make statistic testing based on RSSI distributions, Voiceprint adopts RSSI time series as the vehicular speech and compares the similarity among all received series. Voiceprint does not rely on any predefined radio propagation model, and conducts independent detection without support of the centralized node. Moreover, we improve Voiceprint by allowing it to conduct detection on Service Channel (SCH) to shorten observation time. Furthermore, we extend Voiceprint with change-points detection to identify those illegitimate nodes performing power control. Extensive simulations and real-world experiments demonstrate that Voiceprint is an effective method considering the cost, complexity, and performance.

According to the requirements of informational weapon equipment, the application of polymorphic real-time computation technology in complex embedded system is studied in this paper. The polymorphic real-time computation model for unmanned aerial vehicle(UAV) is designed and verified from the aspects of modal management and polymorphic tasks. The simulation results show that the proposed model can provide modal decision and ensure its real-time calculation to adapt the demand change of UAV system.

With the emerging vehicular network and the possible diverse applications, Intelligent Transportation Systems (ITS) have been evolving to Cooperative ITS (C-ITS, CITS) with connected intelligent vehicles, and the topics in this domain also raise more and more research interesting recently. However, subjecting to the immaturity of V2X communication technology and the deployment of intelligent vehicles in large scale, such studies and the corresponding verification are all challenged within current situations. Focusing on several emerging features, such as cyber-physical fusion, vehicular networking, service-carrier etc, a new simulator QoS-CITS for such service-oriented C-ITS is designed. To enhance the adaptivity, an para-reconfigurable architecture is firstly adopted, within which all reservation-based models of traffic objects, state-driven behaviors, cooperation mechanisms, and policies proposed for service-oriented C-ITS are implemented. And then, a series of experiments are conducted via employing parameters within typical scenes, and all fundamental functions of QoS-CITS are verified. With this simulator, researchers can carry out various experiments, not only classic ones but service-oriented, via setting parameters according to their study and verification requirements, and then analyze the performance with statistics data automatically recorded.

Vehicular Ad Hoc Networks (VANETs) enable vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications that bring many benefits and conveniences to improve the road safety and drive comfort in future transportation systems. Sybil attack is considered one of the most risky threats in VANETs since a Sybil attacker can generate multiple fake identities with false messages to severely impair the normal functions of safety-related applications. In this paper, we propose a novel Sybil attack detection method based on Received Signal Strength Indicator (RSSI), Voiceprint, to conduct a widely applicable, lightweight and full-distributed detection for VANETs. To avoid the inaccurate position estimation according to predefined radio propagation models in previous RSSI-based detection methods, Voiceprint adopts the RSSI time series as the vehicular speech and compares the similarity among all received time series. Voiceprint does not rely on any predefined radio propagation model, and conducts independent detection without the support of the centralized infrastructure. It has more accurate detection rate in different dynamic environments. Extensive simulations and real-world experiments demonstrate that the proposed Voiceprint is an effective method considering the cost, complexity and performance.

As driverless vehicles are increasingly becoming possible, so does the use of such vehicles as intelligent carriers in different domains. Intelligent transportation systems (ITSs) show increasingly heterogeneous, cyber-physical, cooperative, and service-oriented features and are beginning to be merged with the emerging cyber-physical-social systems. Given this new trend, how to make these intelligent vehicles cooperate more safely and efficiently with one another according to novel constraints, such as mission type and quality-of-service (QoS), has become a vital aspect of cooperative ITS (C-ITS). With these emerging characteristics, the classical passing-through-intersection problem has gained new connotations, worth further exploring. After analyzing the essences of this new problem, service-oriented cooperation models and mechanisms for whole autonomous vehicles approaching intersections are investigated in this paper. First, related traffic objects and possible vehicular behaviors are abstracted and modeled with the cyber-physical cooperative features and QoS constraints. A new reservation-based scheduling procedure is then conducted by employing the concepts of vehicle-to-infrastructure communication, and typical vehicular passing-through behaviors and several spatial-temporal constraints are designed to coordinate vehicles passing through an intersection divided as a series of continuous static critical sections. Given these considerations, a priority-based centralized scheduling algorithm, named csPriorFIFO, which adopts a novel priority inheritance mechanism to promote the traffic QoS of emergent vehicles, is proposed. Finally, all these designs are implemented in a traffic simulator named QoS-CITS, and the functions and the performance of these studied methods are verified and compared.

Real-time networks, such as industrial network, field bus and so on, have been becoming one vital component to develop large-scale and cooperative embedded systems. As one important branch, the wireless mode of realtime networks also raises more and more attentions in recent years since its conveniences to construct a flexible control system. Ethernet Powerlink is such a typical real-time industrial network protocol, and provides a master-slave, time-slot mechanism that can well avoid radio collisions. With the designed FPGA-based hardware node, in this paper, a new method to extend wireless capability of Powerlink is explored and described. Concretely, Powerlink architecture and especially its original mechanisms are analyzed at first. For effectively connecting OpenMAC module of Powerlink and RF module, an interface logic at MAC layer is introduced and is typically designed in a multiplexing mode with a dual-FIFO logic according to the limited resource on FPGA, some key designs and mechanisms of which are detailed later. Finally, all designed mechanisms and logics are implemented as an extension part of IP core of Powerlink in VHDL language, and the communication functions and performance of such extended protocol are verified.

Scheduling driverless vehicles with different priorities to pass through intersections efficiently and safely has been becoming an important passing-through intersection (PTI) problem in the field of novel intelligent traffic systems (ITS), which is increasingly becoming cyber-physical-fused and social-service-oriented. Considering new emerging features with possible priorities, a novel centralized priority scheduling mechanism is mainly explored in this paper. First, related pivotal aspects of environment and driverless vehicles are modeled by fusing their physical and kinematic characters. Based on these models, PTI-related motions are further abstracted as several reservation-oriented standard states and actions. Then, an event-triggered and state-driven autonomous control procedure is designed. By mapping vehicular relations in spatiotemporal domain into time-distance windows, a universal passing-through principle, rules, and priority-based scheduling mechanisms are proposed and described in detail. Finally, a priority scheduling algorithm sPriorFIFO is proposed and designed. These models and mechanisms are then implemented within an algorithm simulator, through which scheduling performances are verified and evaluated.

Networked cooperation is one primarily problem for autonomous Cyber-Physical Systems(CPS) that feature cyber-physical fusion, and smart behavior. After analyzing current contributions, a three-level task model, with which missions can be recognized and mapped to tasks, is studied firstly in this paper. And then, based on a designed uniform model of the surroundings, external missions and internal computational resources, two hybrid models, which are autonomously cooperative and real-time reactive by taking advantages of agent and adaptive control, are illuminated respectively. Further, a hierarchal cooperation mechanism of CPS fleet is put forward. With this mechanism, the dynamic topologies of a fleet will be more flexible and dependable, and by adopting an intelligent algorithm, all tasks mapped from a mission can be autonomously assigned to suitable members under decentralized consultation or centralized allocation mode. In this paper, an optimized genetic algorithm is employed to illustrate the process of proposed mechanisms.

Due to the strict requirements of reliabilities in safety-critical domains, this paper researches and proposes an embedded adaptive real-time software fault-tolerant method based on hierarchical and multi-strategy structure. This paper mainly researches on the key issues and appropriate solutions of checkpoints rollback recovery and adaptive task migration. Based on this fault-tolerant method, this paper gives the algorithms of fault monitoring period, checkpoints number and a constraint condition of setting checkpoints. The test result shows that the method improves embedded software reliability effectively under the premise of no additional redundant hardware resources and ensuring the system real-time requirements.

Now, autonomous tasks planning and allocating (TPA) in Multi Agent System (MAS) has been one key and fundamental problem to promote the intelligent level of such system. Autonomous TPA means that, all tasks should be (re)planned and (re)allocated automatically according to the synthesis constraints and the dynamic environment aspects, such as the changing mission, status of each member, and topology, etc. In this article, the formal descriptions of hierarchical tasks and models of logic constraints are studied firstly. And then, some new methods are proposed to evaluate the efficiency of synthesis constraints. Moreover, the key elements, e.g. task allocation vector (TAV), are designed with the theory of genetic algorithm (GA), and a TPA problem can be mapped to the solving model of GA. Based on above, the crossover and mutation operators of GA are optimized with the domain knowledge to perfect the solving efficiency and quality while ensuring the randomicity of evolution. The simulation results show that the solving quality and velocity are improved with studied methods.