In this paper, a two-layer game based joint time and power allocation method for an underlay Energy Harvesting Cognitive Radio (EH-CR) network is proposed. The method first models the interplay between the Primary User (PU) and the Secondary Users (SUs) as a Stackelberg game and then models the interplay among the SUs as a Supermodel game in the underlay EH-CR network. Later, a coefficient for evaluating fair ness is introduced in order to promote fairness among the SUs. Subsequently, the utility function of the primary network and the utility function of the secondary network are defined based on their individual profits. By maximizing the secondary network's utility function, the Supermodel game's Nash Equilibrium (NE) solution is achieved. Then, by substituting the NE solution of the Supermodel game into the utility function of the primary network and then maximizing the utility function of the primary network, the NE solution of the Stackelberg game is obtained. Finally, a deterministic strategy can be obtained, which is the time coefficient of equalized spectrum sensing and the equalized power allocation scheme instead of a probabilistic strategy. Simulation outcomes demonstrate that, under the condition of maintaining the communication quality of the PU, the PU's revenue when PH0 = 0.8 can be improved by 18.2% and when PH0 = 0.6 can be improved by 13.3% compared with the conventional method.

The rapid advancement of Industry 5.0 has accelerated the adoption of the Industrial Internet of Things (IIoT). However, challenges such as data privacy breaches, malicious attacks, and the absence of trustworthy mechanisms continue to hinder its secure and efficient operation. To overcome these issues, this paper proposes an enhanced blockchain-based data storage framework and systematically improves the Delegated Proof of Stake (DPoS) consensus mechanism. A four-party evolutionary game model is developed, involving agent nodes, voting nodes, malicious nodes, and supervisory nodes, to comprehensively analyze the dynamic effects of key factors—including bribery intensity, malicious costs, supervision, and reputation mechanisms—on system stability. Furthermore, novel incentive and punishment strategies are introduced to foster node collaboration and suppress malicious behaviors. The simulation results show that the improved DPoS mechanism achieves significant enhancements across multiple performance dimensions. Under high-load conditions, the system increases transaction throughput by approximately 5%, reduces consensus latency, and maintains stable operation even as the network scale expands. In adversarial scenarios, the double-spending attack success rate decreases to about 2.6%, indicating strengthened security resilience. In addition, the convergence of strategy evolution is notably accelerated, enabling the system to reach cooperative and stable states more efficiently. These results demonstrate that the proposed mechanism effectively improves the efficiency, security, and dynamic stability of IIoT data storage systems, providing strong support for reliable operation in complex industrial environments.

This paper proposes a novel Intelligent Reflecting Surface (IRS)-assisted interweave Cognitive Internet of Vehicles (CIoV) network under malicious jamming attacks, where the IRS enhances communication performance by establishing additional links. In order to maximize the sum transmission rate of Vehicle-to-Infrastructure (V2I) links, we propose an optimization problem that jointly optimizes wireless resource allocation, such as spectrum and transmit power for Vehicle Users (VUs) and IRS phase shift. Because this problem is non-convex and complicated, we further propose a Heterogeneous Multi-agent Transformer-enhanced Dueling Double Deep Q-Network (HMA-TD3QN) based resource allocation method, where VUs and Secondary Base Station (SBS) act as distinct heterogeneous agents can independently perform resource allocation and phase shift optimization. The Transformer neural network architecture can better adapt to long sequence input states and extract relevant features from complex input states through the attention mechanism. Simulation results indicate that the proposed HMA-TD3QN method achieves improvements of 24.42%, 20.79%, and 22.25% over the basic HMA-DQN under three different jamming strategies, highlighting the effectiveness of IRS technology in enhancing the Quality of Service (QoS) and jamming resilience of CIoV network.

The explosive growth of Internet of Things (IoT) data demands secure and reliable storage, where traditional centralized solutions often fall short. Blockchain offers decentralization and tamper-resistance, making it a promising foundation for IoT. However, IoT blockchain systems based on Delegated Proof-of-Stake (DPoS) face challenges such as weak node incentives, unfair reward distribution, and low consensus efficiency. This paper proposes a fairness-aware incentive mechanism that accounts for both node capability and effort under information asymmetry. By incorporating fairness preferences into the contract design, the mechanism improves participation and motivates sustained effort. Theoretical analysis and simulation results show that our approach enhances throughput by about 15%, while achieving revenue fairness, incentive compatibility, and stronger consensus performance. The mechanism’s adaptability makes it suitable for diverse IoT application scenarios.

Cognitive Internet of Vehicles (CIoV) adds the cognitive engine based on traditional Internet of Vehicles (IoV), which can improve spectrum utilization. However, spectrum sensing data falsification (SSDF) attacks pose a threat to CIoV network security. To ensure the full utilization of spectrum resources and protect primary users transmission, this article combines blockchain with CIoV to defend against SSDF attacks in the presence of vehicle users (VUs) entering and leaving the network. Specifically, this article introduces a virtual currency called Sencoins serve as credential for VUs to purchase transmission shares. And this article proposes a reward and punishment mechanism and a hybrid Proof-of-Stake (PoS) and Proof-of-Work (PoW) mining model to thwart the motivation of the VUs to launch SSDF attacks. On this basis, this article investigates the dynamics of SSDF attack strategy choice of VUs, and uses the largest Lyapunov exponent (LLE) to determine the critical value of Sencoins that avoids the system to exhibit chaotic behavior. To describe the uncertainty of the population proportion of VUs that choose different attack strategies due to high-speed movement and the VUs entering and leaving the CIoV network, this article introduces Gaussian white noise into the replication dynamics equation and builds the Itô stochastic evolutionary game model, and solves it according to the stability judgment theorem of stochastic differential equations and stochastic Taylor expansion. Finally, simulation results verify that the proposed method can quickly and effectively thwart SSDF attacks in the CIoV network. And compared with traditional methods, the proposed method can improve the efficiency of defending against SSDF attacks by 567% and the average throughput by 25%.

This article proposes a novel graph recurrent neural network (GRNN)-based approach for detecting the eavesdropping attacks in smart grid wireless communication systems enabled by simultaneous wireless information and power transfer (SWIPT). By leveraging the graph-centric nature of GRNNs, the proposed method effectively learns the topological structure and the edge features of the wireless sensor networks (WSNs), enabling the detection of the eavesdropping attacks in dynamic WSNs. This article mathematically models the channel state information (CSI) under the man-in-the-middle eavesdropping attacks based on the physical-layer security (PLS) in SWIPT networks. Moreover, this article sets up a real-world testbed to create the training and testing data sets. The proposed GRNN model can handle large-scale complex topologies and dynamic eavesdropping networks, accurately detect eavesdropping behaviors, and enhance the security of information transmission in WSNs. Simulation results demonstrate that, compared with the algorithms, such as support vector machine (SVM), K-nearest neighbors (KNNs), convolutional neural network (CNN), graph convolutional network (GCN), and gated recurrent unit (GRU), the proposed method exhibits stronger robustness under complex attack scenarios, achieving a detection accuracy of over 95%. This article provides a novel and effective graph learning solution for the smart grid wireless communication security, which is of great significance to ensure the stable and reliable operation of the smart grids.

Cognitive radio (CR) is regarded as the key technology of the 6th-Generation (6G) wireless network. Because 6G CR networks are anticipated to offer worldwide coverage, increase cost efficiency, enhance spectrum utilization, and improve device intelligence and network safety. This article studies the secrecy communication in an energy-harvesting (EH)-enabled Cognitive Internet of Things (EH-CIoT) network with a cooperative jammer. The secondary transmitters (STs) and the jammer first harvest the energy from the received radio frequency (RF) signals in the EH phase. Then, in the subsequent wireless information transfer (WIT) phase, the STs transmit secrecy information to their intended receivers in the presence of eavesdroppers while the jammer sends the jamming signal to confuse the eavesdroppers. To evaluate the system secrecy performance, we derive the instantaneous secrecy rate and the closed-form expression of secrecy outage probability (SOP). Furthermore, we propose a deep reinforcement learning (DRL)-based framework for the joint EH time and transmission power allocation problems. Specifically, a pair of ST and jammer over each time block is modeled as an agent which is dynamically interacting with the environment by the state, action, and reward mechanisms. To better find the optimal solutions to the proposed problems, the long short-term memory (LSTM) network and the generative adversarial networks (GANs) are combined with the classical DRL algorithm. The simulation results show that our proposed method is highly effective in maximizing the secrecy rate while minimizing the SOP compared with other existing schemes.

Cognitive Industrial Internet of Things (CIIoT) permits Secondary Users (SUs) to use the spectrum bands owned by Primary Users (PUs) opportunistically. However, in the absence of the PUs, the selfish SUs could mislead the normal SUs to leave the spectrum bands by initiating a Primary User Emulation Attack (PUEA). In addition, the application of Energy Harvesting (EH) technology can exacerbate the threat of security. Because the energy cost of initiating a PUEA is offset to some extent by EH technology which can proactively replenish the energy of the selfish nodes. Thus, EH technology can increase the motivation of the selfish SUs to initiate a PUEA. To address the higher motivation of the selfish SUs attacking in CIIoT scenario where the EH technology is applied, in this paper, an EH-PUEA system model is first established to study the security countermeasures in this severe scenario of PUEA problems. Next, a new reward and punishment defense management mechanism is proposed, and then the dynamics of the selfish SUs and the normal SUs in a CIIoT network are studied based on Evolutionary Game Theory (EGT), and the punishment parameter is adjusted according to the dynamics of the selfish SUs to reduce the proportion of the selfish SUs’ group choosing an attack strategy, so as to increase the throughput achieved by the normal SUs’ group. Finally, the simulation results show that the proposed mechanism is superior to the conventional mechanism in terms of throughput achieved by the normal SUs’ group in CIIoT scenario with EH technology applied.

In cognitive radios, energy detector is considered for spectrum sensing in the literature. However, its performance deteriorates rapidly if the noise power is not known exactly. Moreover, due to the presence of a colored channel interferer or some other reasons, the conventional white Gaussian noise may become colored. In order to solve these problems, this paper proposes several multicumulant based spectrum sensing methods: generalized likelihood ratio test (GLRT) based multicumulant (GLRTMC) based detection method and multiantenna-assisted multicumulant (MAMC) based detection method. GLRTMC detection method is derived from generalized likelihood ratio test and assumed to be near optimum in theory. MAMC detection method, on the other hand, by using multiple antennas, is a complexity-reduced detector and allows us to make a compromise between performance and complexity. It is well known that cumulants higher than second order are zero for Gaussian distributions. Thus, GLRTMC detection method and MAMC detection method can extract a non-Gaussian signal from Gaussian noise even when the noise is colored. In addition, the proposed methods are nonparametric in the sense that they do not require any exact prior knowledge about the signal or the noise, such as noise power or cyclic frequencies. Hence they are immune from noise uncertainty. Simulation experiments are provided to show the validity and the superiority over single-cumulant based detector of the proposed multicumulant based detectors.

In cognitive radios, energy detector is considered for spectrum sensing in the literature. However, its performance deteriorates rapidly when noise power is fluctuating. In order to solve this problem, this paper proposes a spectrum sensing method based on multiple cumulants. Although cumulant-based detection methods have been extensively studied, they usually consider a single cumulant of a certain order and a certain lag. Therefore, they only make limited use of the rich statistical information contained in the signal. The proposed detection method is based on multiple cumulants of arbitrary orders and arbitrary lags hence it is superior and more flexible. In addition, the proposed detector is nonparametric in the sense that it does not require any exact prior knowledge about the signal or the noise, such as noise power or cyclic frequencies. Thus it is immune from noise uncertainty. Simulation experiments are provided to show the validity and the superiority over single-cumulant based detector of the proposed multicumulant based detector.