The secure deployment of Federated Learning (FL) is critically undermined by statistical data heterogeneity and a profound vulnerability to adversarial attacks, these weaknesses are exacerbated by FL’s privacy-preserving preclusion of large-scale, centralized data for robust training. Existing proto-typebased methods suffer from representation collapse when naively aggregating from non-IID clients, while generative approaches often lack a principled mechanism for synthesizing features that confer adversarial resilience. We introduce Federated Contrastive Diffusion Prototypes (Fed-CDP), a novel paradigm that transforms the server from a passive aggregator into an active synthesis hub for robust features. Fed-CDP aggregates lightweight client prototypes to serve as semantic anchors, guiding a server-side diffusion model via a contrastive objective. This process synthesizes a high-fidelity feature space explicitly optimized for maximal inter-class separability, a property intrinsically linked to robust generalization. These server-generated features are then distributed to clients as a potent regularizer, aligning disparate local models and directly mitigating client drift. Our extensive evaluations across multiple challenging datasets establish that Fed-CDP outperforms existing state-of-the-art baselines. For instance, on CIFAR-100 under severe heterogeneity (α = 0.1), Fed-CDP surpasses leading methods by nearly 5% in standard accuracy and over 9% in robust accuracy under Projected Gradient Descent attacks. Fed-CDP provides a new blueprint for building secure and high-performance collaborative AI, laying the foundation for trustworthy systems in critical sectors like finance and multi-institutional healthcare.

Quality control of sensing data poses a fundamental challenge in Spatial Crowdsourcing (SC) applications, and various incentive mechanisms have been proposed to effectively motivate workers. However, real-world SC systems involve multiple stakeholders, namely workers, platforms, and requesters, who exhibit bounded rationality. Moreover, prior research has often neglected the practical aspect of platforms needing to verify sensing data against ground truth information, which incurs a cost. Therefore, platforms face the constraint of verification cost and must judiciously select the verification rate. In this paper, a S ustainable Q uality C ontrol S ystem (SQCS) is proposed, which employs a three-party evolutionary game to model the dynamic interaction among workers, platforms, and requesters in practical SC scenarios. Based on this dynamic model: (1) In theory, we derive the minimum verification rate required for effective worker monitoring and the maximum verification cost that platforms can sustainably bear via the analysis of evolutionary stable strategies. These findings serve as guidelines for achieving sustainable quality control within the SQCS framework; (2) In practice, we apply the SQCS framework to the SC scenario of urban air pollution monitoring. The optimal quality control strategy that maximizes overall social benefits is analyzed via numerical methods . Extensive experimental validation is conducted to demonstrate the effectiveness of the proposed theory and practical approach. These findings provide valuable insights for the sustainable development of SC systems.

The multiuser cooperative offloading mobile edge computing (COMEC) system has attracted much attention because it can realize delay-sensitive tasks. However, in the coupling optimization of offloading decision and resource allocation, the existing numerical optimization algorithms are difficult to obtain high-quality optimization solutions. In this paper, we propose a differential dynamic gradient descent (DDGD) optimization algorithm to solve the above optimization problems. DDGD algorithm decomposes the constrained NP-hard optimization problem into two network layers and integrates the constraint function into a larger end-to-end training network. These two-layer networks encode the dependencies and optimization constraints between parameter hidden states, which cannot be captured by a numerical optimization model or a full connection layer neural network. Because the ring learning and self-repeating learning architecture are adopted and the information is stored in the differential dynamics network, the proposed algorithm can achieve better and more intelligent decision-making in searching the solution trajectory without setting accurate parameters in advance and reduce the complexity of the network. We show that compared with the baseline method, the DDGD method has superior optimization performance in the energy consumption optimization of COMEC.

Recent developments in collection, computation and communication have expanded the way of data reporting in intelligent Internet of Things (IoT). However, diversity and complexity of data sources also impose new trust challenge in data collection process since untrust reporters tend to report false or even malicious data, which highlights the need to develop a novel methodology to solve such challenge. Thus, based on this domain, inspired by deterrence theory, this article proposes a blockchain-based trust system with assistant of drones to deter malicious data reporting in intelligent IoT. Specifically, to deter malicious data reporting, based on the blockchain technology, the data sensed by fully trusted drones is public published on blockchain showing participants the data standards, named as malicious deterrence scheme. This scheme provides a barrier for malicious reporters to arbitrarily publish false data to blockchain, since the false data can be easily detected while they cannot deny. Second, to further reduce malicious data reporting, a strict penalty mechanism is proposed to punish malicious reporters who have reported false data to blockchain to reduce the malicious data reporting in the following task through punishment. Third, note that the sensing of data standard generates additional costs, therefore, a drone flight route scheme based on a simper deep reinforcement learning with multihead attention mechanism (MA-DRL) is designed to reduce the flight distance for drones. Finally, extensive experiments demonstrate efficiency of our proposed system in terms of reducing malicious data reporting in advance as well as reducing drone flight distance.

Contemporarily, the fast development of computing, communication, and storage technology has revolutionized the way that various data-based applications reach massive data from underlying sensor networks. However, such a process also raises two challenging but critical issues: 1) trustworthy and 2) privacy issue for data collectors. Therefore, this article proposes a novel system, which is designed over the distributed mobile-edge network to sufficiently exploit advantages of blockchain and differential privacy (DP) to collect trustworthy data and protect privacy for data collectors. First, to improve trustworthiness of data collections, a new consensus mechanism is proposed for blockchain-based data collection structure, which comprehensively incorporates trustworthy, collection contribution, and throughput together to prefer data collectors for the next block. Second, with the assistance of fully trusted devices, a verifiable trustworthy evaluation strategy is designed to accurately compute the trustworthiness for data collectors. Third, we enforce DP on the data stored in a global blockchain maintained by the cloud server to protect privacy for data collectors without influencing data availability. Finally, both theoretical analyses and experimental results prove that the proposed system comprehensively improves performance of data collections in distributed network without adding any additional cost for the cloud server, compared to other schemes.

Recently, rapid deployment on the fifth-generation (5G) networks has brought great opportunities for enabling data-intensive applications and brings an extending expectation on the developments of 6G. A basic requirement to develop 6G networks is to reach data with low latency, low cost, and high coverage in smart Internet of Things (IoT). Therefore, this article proposes a novel machine learning-based approach to collect data from multiple sensor devices by cooperation between vehicle and unmanned aerial vehicle (UAV) in IoT. First, a genetic algorithm is utilized to select vehicular collectors to collect massive data from sensor devices, which aims to maximize coverage ratio and to minimize employment cost. Second, we design a novel deep reinforcement learning (DRL)-based route policy to plan collection routes of UAVs with constrain energy, which simplifies the network model, accelerates training speeds, and realizes dynamic planning of flight paths. The optimal collection route of a UAV is a series of outputs based on the proposed DRL-based route policy. Finally, our extensive experiments demonstrate that the proposed scheme can comprehensively improve the coverage ratio of massive data collections and reduce collection costs in smart IoT for the future 6G networks.

A large number of devices with communication and sensing capabilities are connected to the 5G network and facilitate distributed industrial Internet of Things (IIoT) applications. Those devices can update their program code by using software-defined technologies and upgrade functions without hardware replacement, thus, greatly facilitating the development of IIoT applications. Disseminating code safely to these devices is a pivotal issue. However, the existing methods rarely consider whether the code received by devices is integrated. Therefore, an unmanned aerial vehicle (UAV) assisted trustworthy code dissemination (UTCD) framework in 5G-enabled intelligent IIoT systems is proposed to select credible mobile vehicles (MVs) to disseminate code with opportunistic routing style. In the UTCD framework, a verifiable trust evaluation scheme is proposed to identify the trust of the MVs by sending the UAV to collect the code wait to be verified (CWV) directly from the selected devices (bedrock devices) as evidence. This scheme is a fundamental change from the previous passive, indirect and unverifiable trust evaluation schemes. In UTCD, a bedrock devices selection scheme is presented by selecting as few bedrock devices as possible to collect as much the CWV as possible with the minimum cost. After that, based on the CWV, which is collected from bedrock devices, a complete trust evaluation scheme is proposed to obtain the trust of MVs. Furthermore, a UAV trajectory optimization algorithm is proposed to obtain as much CWV as possible within the limited conditions of UAV. Finally, comprehensive experiments conduct on a real-life dataset demonstrate that the proposed scheme outperforms the existing schemes in terms of the efficiency and security of code dissemination.

Large number of Internet of Things (IoT) applications with intelligent sensing devices (ISDs) are combined with cloud computing to collect and process data more efficiently. However, ISDs are vulnerable to various attacks. Compromised devices may maliciously provide unreliable data to the cloud, causing damage to IoT applications. Therefore, it is a critical issue to design an effective mechanism to ensure the security of data collection in cloud computing. In this article, a cloud-assisted reliable trust computing (CRTC) scheme is proposed to identify the trust of ISDs at low cost, providing high-quality data for IoT applications. The CRTC scheme mainly includes the following parts: First, a reliable approach of obtaining the real data of ISDs at a low cost is proposed to identify the trust of ISDs. In the proposed method, ISDs fits the forwarding packets information to form inspection information (II) with a small amount of data and then sends II to inspection nodes, effectively reducing the cost of routing II. Second, an effective method of trust computing is given to evaluate the trustworthiness of ISDs based on the data and II collected by unmanned aerial vehicles (UAV). Then, the aggregators are selected from high-trust ISDs to ensure secure data collection. Third, to obtain more reliable trust at lower cost, a trajectory optimization algorithm for UAV is proposed to collect as much II as possible and reduce the moving distance. Theoretical analysis and experimental results show that the proposed CRTC scheme is superior to previous strategies in terms of the success rate of data collection, the speed of identifying the trust of ISDs, and the UAV's trajectory distance.

Billions of devices are connected to the Internet of Things (IoT). These devices generate a large volume of data, which poses an enormous burden on conventional networking infrastructures. As an effective computing model, edge computing is collaborative with cloud computing by moving part intensive computation and storage resources to edge devices, thus optimizing the network latency and energy consumption. Meanwhile, the software-defined networks (SDNs) technology is promising in improving the quality of service (QoS) for complex IoT-driven applications. However, building SDN-based computing platform faces great challenges, making it difficult for the current computing models to meet the low-latency, high-complexity, and high-reliability requirements of emerging applications. Therefore, a cloud-mobile edge computing (MEC) collaborative task offloading scheme with service orchestration (CTOSO) is proposed in this article. First, the CTOSO scheme models the computational consumption, communication consumption, and latency of task offloading and implements differentiated offloading decisions for tasks with different resource demand and delay sensitivity. What is more, the CTOSO scheme introduces orchestrating data as services (ODaS) mechanism based on the SDN technology. The collected metadata are orchestrated as high-quality services by MEC servers, which greatly reduces the network load caused by uploading resources to the cloud on the one hand, and on the other hand, the data processing is completed at the edge layer as much as possible, which achieves the load balancing and also reduces the risk of data leakage. The experimental results demonstrate that compared to the random decision-based task offloading scheme and the maximum cache-based task offloading scheme, the CTOSO scheme reduces delay by approximately 73.82%-74.34% and energy consumption by 10.71%-13.73%.

Reducing energy consumption, increasing network throughput, and reducing delay are the pivot issues for wake-up radio- (WuR-) enabled wireless sensor networks (WSNs). In this paper, a relay selection joint consecutive packet routing (RS-CPR) scheme is proposed to reduce channel competition conflicts and energy consumption, increase network throughput, and then reduce end-to-end delay in data transmission for WuR-enabled WSNs. The main innovations of the RS-CPR scheme are as follows: (1) Relay selection: when selecting a relay node for routing, the sender will select the node with the highest evaluation weight from its forwarding node set (FNS). The weight of the node is weighted by the distance from the node to sink, the number of packets in the queue, and the residual energy of the node. (2) The node sends consecutive packets once it accesses the channel successfully, and it gives up the channel after sending all packets. Nodes that fail the competition sleep during the consecutive packet transmission of the winner to reduce collisions and energy consumption. (3) Every node sets two thresholds: the packet queue length threshold Nt and the packet maximum waiting time threshold Tt . When the corresponding value of the node is greater than the threshold, the node begins to contend for the channel. Besides, to make full use of energy and reduce delay, the threshold of nodes which are far from sink is small while that of nodes which are close to sink is large. In such a way, nodes in RS-CPR scheme will select those with much residual energy, a large number of packets, and a short distance from sink as relay nodes. As a result, the probability that a node with no packets to transmit becomes a relay is very small, and the probability that a node with many data packets in the queue becomes a relay is large. In this strategy, only a few nodes in routing need to contend for the channel to send packets, thereby reducing channel contention conflicts. Since the relay node has a large number of data packets, it can send many packets continuously after a successful competition. It also reduces the spending of channel competition and improves the network throughput. In summary, RS-CPR scheme combines the selection of relay nodes with consecutive packet routing strategy, which greatly improves the performance of the network. As is shown in our theoretical analysis and experimental results, compared with the receiver-initiated consecutive packet transmission WuR (RI-CPT-WuR) scheme and RI-WuR protocol, the RS-CPR scheme reduces end-to-end delay by 45.92% and 65.99%, respectively, and reduces channel collisions by 51.92% and 76.41%. Besides, it reduces energy consumption by 61.24% and 70.40%. At the same time, RS-CPR scheme improves network throughput by 47.37% and 75.02%.