The purpose of this study is to investigate the air quality levels of Nanyang city according to Chinese air quality standards. Therefore, in this study, fine particulate matter (PM_2.5), coarse particulate matter (PM₁₀), and total suspended particulate (TSP) were analyzed from 19 November 2021 to 19 March 2022 in Nanyang city. The results show that the average concentrations of PM_2.5, PM_10, and TSP were 106.47 µg/m³, 137.32 µg/m³, and 283.40 µg/m³, respectively. The numbers of days that meet the national secondary air quality standard of 24-h average concentrations were 29.75% for PM_2.5, 63.64% for PM₁₀, and 63.64% for TSP, indicating that most of the time, the air quality of Nanyang city remains polluted in winter, especially with more contributions of PM_2.5 compared to PM₁₀ and TSP. The higher concentrations were observed between 07:00 and 08:00, suggesting that vehicular emissions can be a major cause of air pollution in Nanyang city. The results also show a significant positive correlation between particulate matter and relative humidity, and a weak correlation with temperature and wind speed, which suggests that higher relative humidity increases the formation of particulate matter. This study can provide theoretical support for the local government to formulate air pollution prevention and control policies for Nanyang city.

Traffic identification is a key task for any Internet Service Provider (ISP) or network administrator. Machine learning method is an important researchmethod on traffic identification, while impact of the asymmetry router on the traffic identification is considered, so this paper analyzes the impact of asymmetry routing on traffic identification, and proposes an effective method to decrease the impact, and experimental results show the auto adaptive algorithm can improve the traffic identification.

In this work, we focus on the issue of secure communication in energy harvesting untrusted relay networks taking into account the impact of hardware impairments, where an energy‐constrained relay, powered by received radio frequency signals, attempts to decode sensitive data from the source without authorization. To achieve secure communication, we employ destination‐assisted jamming to prevent the untrusted relay from decoding sensitive data while serving as a useful energy source. We derive the interception probability (IP), connection outage probability (COP), and effective secrecy throughput (EST) in closed form for time switching relaying (TSR) strategy. Subsequently, we conduct an asymptotic analysis in the high signal‐to‐noise ratio (SNR) regime to provide practical insights into how various system parameters affect security, reliability, and efficiency, including time switching factor and the level of hardware impairments. Our findings demonstrate that the presence of hardware impairments improves security performance but compromises reliability performance, which does more harm than good.

Ultra-reliable low-latency communication (URLLC) presents significant challenges in simultaneously guaranteeing stringent latency bounds, ultra-high reliability, and efficient resource utilization under dynamic channel conditions. To address these joint constraints, a novel framework that integrates a reconfigurable intelligent surface (RIS) with cooperative non-orthogonal multiple access (NOMA) is proposed for short-packet communications. Two distinct phase configuration designs for the RIS are considered, i.e., a near-user priority strategy (NUPS) and a far-user priority strategy (FUPS). The NUPS configures the RIS to enhance the received signal power for the near user, while the FUPS optimizes the phase shifts to maximize the received power for the far user. Closed-form expressions that characterize the average block error rate (BLER) of the near and far users under the two proposed strategies in the presence of hardware impairments are derived. Specifically, the analysis for the far user considers both selection combining (SC) and maximum ratio combining (MRC) reception schemes. Based on the average BLER, we then derive a closed-form expression for the effective throughput. Simulation findings reveal the following: (1) The far user in the proposed cooperative NOMA achieves a lower average BLER than in the non-cooperative NOMA. (2) When the RIS is deployed in close proximity to the base station (BS), the NUPS can effectively leverage the RIS to enhance the far user’s signal quality through cooperation, without sacrificing the near user’s priority; and (3) SC serves as a low-complexity alternative that achieves near-optimal performance when inter-user channel conditions are favorable.

The heterogeneity of data between different clients limits the performance of federated learning training, and hyper-knowledge distillation has emerged as a potential solution to address heterogeneous data. However, existing hyper-knowledge distillation methods often sacrifice efficiency to improve model performance. To address these challenges, this paper proposes FedHDE, a novel federated learning framework based on hyper-knowledge distillation and efficient tensor optimization. FedHDE integrates a Dynamic Client Selection strategy within a teacher–assistant–student hierarchical structure to enhance global aggregation accuracy and reliability without relying on public datasets or server-side models. Furthermore, an efficient tensor optimization algorithm is introduced to improve numerical stability and reduce computation time. Theoretical analysis demonstrates that FedHDE achieves a convergence rate of $$O(1/\sqrt{T})$$ O ( 1 / T ) , ensuring efficient and stable learning. Experimental results on the CIFAR10 dataset show that FedHDE improves both local and global accuracy by 0.14%–4.32%, while reducing training time by 65.28% compared to state-of-the-art baselines.Code is public at https://anonymous.4open.science/r/FedHDE .