In non-terrestrial networks (NTN), passive Internet of Things (IoT) systems typically face challenges including unstable energy supply, limited signal coverage, and complex channel environments. This paper proposes an innovative optimization framework that jointly designs the trajectory and phase configuration of unmanned aerial vehicle (UAV) mounted reconfigurable intelligent surfaces (RIS) based on trust region policy optimization (TRPO) algorithms. The framework encompasses two key technical innovations: first, a transmission protocol based on orthogonal time-slot allocation that maximizes backscatter communication performance while ensuring efficient energy harvesting; second, a position-phase decoupled optimization strategy that employs a weighted geometric center method to determine UAV position, followed by TRPO optimization of RIS phase configuration, effectively overcoming learning convergence difficulties caused by position-dependent channel variations. Simulation results demonstrate that our scheme exhibits excellent performance across various scenarios, providing an efficient and practical solution for passive IoT applications in dynamic NTN.

Federated learning (FL) enables collaborative training of machine learning models across distributed medical data sources without compromising privacy. However, applying FL to medical image analysis presents challenges like high communication overhead and data heterogeneity. This paper proposes novel FL techniques using explainable artificial intelligence (XAI) for efficient, accurate, and trustworthy analysis. A heterogeneity-aware causal learning approach selectively sparsifies model weights based on their causal contributions, significantly reducing communication requirements while retaining performance and improving interpretability. Furthermore, blockchain provides decentralized quality assessment of client datasets. The assessment scores adjust aggregation weights so higher-quality data has more influence during training, improving model generalization. Comprehensive experiments show our XAI-integrated FL framework enhances efficiency, accuracy and interpretability. The causal learning method decreases communication overhead while maintaining segmentation accuracy. The blockchain-based data valuation mitigates issues from low-quality local datasets. Our framework provides essential model explanations and trust mechanisms, making FL viable for clinical adoption in medical image analysis.

In this article, a novel system based on the simultaneous iterative reconstructive technique (SIRT) and multiagent deep reinforcement learning is proposed for detection of anomalous geological structures in coal mines. The system employs the SIRT optimization inversion method to construct a computational model for channel wave signal imaging. Then, the back projection technique (BPT) was introduced to the system. By utilizing the BPT algorithm to provide initial values for the SIRT, the channel wave signals can be prescreened, improving the ability of the SIRT algorithm to suppress model noise and enhancing its resolution. Furthermore, we employ multiagent reinforcement learning method for image feature classification of anomalous geological structures. Moreover, we conduct two-dimensional and three-dimensional imaging of four types of changes and energy fluctuations. The results demonstrate a high degree of concordance between the computed channel wave results and the slowness of the measured channel wave signals. Experimental findings validate the exceptional computational accuracy of this novel system, with relative errors and coefficient of deviation both within 1%, surpassing traditional SIRT inversion methods, damped least-squares methods, conjugate gradient methods, and classical algebraic reconstruction methods. These discoveries demonstrate the feasibility and superiority of utilizing transmission tomography imaging technology for the detection of anomalous structures in coal seams, offering new perspectives for underground exploration in coal mines.

The space–air–ground-integrated network (SAGIN) will greatly promote the development of the Internet of Things (IoT). Green IoT will be an important part of SAGIN. Ambient backscatter communication (AmBC) is a potential solution for green SAGIN IoT. To improve the achievable sum rate (ASR) of the AmBC system, we propose a reconfigurable intelligent surface (RIS)-assisted AmBC system. In the single-backscatter device (BD) AmBC scenario, we first give the phase shifts that maximize the gain of the reflection link of the AmBC system, and then give the optimal reflection coefficient. The proposed scheme does not need to solve the convex semidefinite program (SDP) problem and has the characteristics of low computational complexity. In the multi-BD AmBC scenario, we first propose a multi-BD phase shifts initialization strategy to ensure the stability of the proposed scheme. Then, we give the optimal reflection coefficient and phase shifts based on the iterative method. Simulations show that the RIS-assisted AmBC scheme is superior to the non-RIS-assisted AmBC scheme.

This paper looks into the realization of digital twin (DT) technology in unmanned aerial vehicle (UAV) networks. We propose, in particular, a framework for DT-based UAV applications in which distinct jobs in the digital twin interact with UAVs in the physical world via task manager scheduling. Furthermore, we investigate the use of 3D mmWave Radar imaging on UAVs and apply it to the process of radio frequency (RF) characterizing. After that, the method of 3D ray-tracing is employed to accomplish channel modelling of UAVs, which reflects RF domain digital twin match. Finally, we present numerical results to demonstrate that our developed digital twin platform can provide accurate RF presentation of UAV and therefore accomplish smart operating and administration of the actual UAV network.

Person reidentification (re-ID), which is a significant and potential application in the Internet of Things (IoT), aims to retrieve pedestrians of interest given a labeled image in a camera network. Now, it is still existing many challenges that severely influence feature representation in practical scenarios. Many methods adopt the attention mechanism in convolutional neural network (CNN) to improve the ability of feature learning. Although they only apply 1-D attention block in the popular deep learning architecture, the learned features are not discriminative for the feature representation. In this work, we investigate a self-focusing network (SFNet) that considers both the channel-dimensional attention and spatial-dimensional attention to adaptively learn more discriminative features. Namely, we embed the new attention module into the common backbone network, which can focus on the salient region by inhibiting the redundant features. Specifically, we design eight variants of the channel-dimensional attention and spatial-dimensional attention throughout the entire network and explore the most powerful feature representation. The heatmaps of different layers are visualized to intuitively present the performance of SFNet. Furthermore, we compare SFNet with the prior work on three popular person re-ID benchmarks by abundant experiments.

The uncontrollability of the radio frequency (RF) environment is one of the main obstacles hindering the popularization of ambient backscatter communication devices, because the devices need strong signal to maintain the overhead of backscatter circuits and the reflection of the modulated signals. The intelligent reflector (IRS) can improve the radio frequency environment by adjusting the phase and amplitude of the incident signal, which provides the possibility for the widespread deployment of ambient backscatter communication devices. In this paper, we introduce a novel IRS-assisted ambient backscatter communications system (ABCS), in which the signal of ABCS rides on the signal of the primary system. The two systems share the same receiver, and the signals of the two systems can be demodulated separately based on continuous interference cancellation (SIC) technology. The purpose of this paper is to design the beamforming vector and IRS phase shift jointly to minimize the AP's transmit power while ensuring the quality of service of the ABCS and the primary communication system. Due to the non-convex nature of the problem, the time complexity of solving the problem through exhaustive search will be very high. Therefore, we propose an iterative-based beamforming vector and IRS phase shift joint design method to minimize the AP transmit power. This method can effectively reduce the transmission power of the access point, and the simulation results prove the effectiveness of the method.

Person reidentification (re-ID), which aims at recognizing the pedestrians captured by multiple nonoverlapping cameras, has attracted more interest due to its significant and potential application in the Internet of Things like intelligent visual surveillance. However, person reID is still a challenging problem in the situations of various pose, similar appearances, partial occlusion, etc. To handle these obstacles, in this article, we investigate an innovative exciting-inhibition network (EINet) that is a two-branch network composed of the exciting branch and the inhibition branch. The channel-spatial attention block that recalibrates the relationship between channels and highlights features at different spatial positions is used in the exciting branch. A novel Soft Batch DropBlock that randomly selects a continuous region of the intermediate feature maps at the same location is applied in the inhibition branch to inhibit the trivial by an inhibitive mask and reinforce learning the remaining regions. We integrate the comprehensive features from both branches for evaluation and show the performance of EINet intuitively using the visualization method. Abundant experiments demonstrate the state-of-the-art performance by comparing with the previous methods on three popular person re-ID benchmarks. For example, our method obtains 95.64% Rank-1 and 88.75% mean average precision (mAP) on Market-1501, and 77.00% Rank-1 and 74.51% mAP on CUHK03-Detect in the single query mode, respectively.

Ambient backscatter communication (AmBC) is one of the candidate solutions for the 6G green internet of things (IoT) network. However, the uncontrollability of the radio frequency (RF) environment is one of the main obstacles hindering the popularization of AmBC. The intelligent reflective surface (IRS) can improve the radio frequency environment by adjusting the phase and amplitude of the incident signal, which provides the possibility for the widespread deployment of AmBC. Currently, there is no discussion about the joint optimization of AmBC and IRS. In this paper, we introduce a novel IRS and AmBC joint design method. The purpose of this method is to jointly design the beamforming vector, the IRS phase shift, and the reflection coefficient of AmBC to minimize the AP’s transmit power while ensuring the quality of service of the AmBC system and the primary communication system. Due to the nonconvexity of the problem, the time complexity of solving the problem through exhaustive search will be very high. Therefore, we propose a joint design method based on an iterative beamforming vector, IRS phase shift, and reflection coefficient to minimize the AP’s transmit power. This method can effectively reduce the transmission power of the access point (AP), and the simulation results prove the effectiveness of the method.

This paper considers the downlink ultra-dense cloud radio access network (C-RAN), which employs multiple radio remote head (RRH) cooperation to guarantee the minimum achievable transmission rate for each user equipment (UE). However, due to the limited orthogonal frequency resources, it is difficult to achieve this goal. To maximize the coverage probability of the system, we focus on the joint user-centric clustering and frequency allocation problem. To reduce the computational complexity, this problem is split into two sub-problems: user-centric clustering and frequency allocation. Firstly, we propose a novel binary user-centric clustering strategy, which includes serving clusters and silent clusters. This strategy determines the acceptable combination of serving clusters and silent clusters to guarantee the minimum transmission rate for each UE and simplify the complexity of the subsequent frequency allocation. Then based on the generated clusters, a new graph generation method is proposed. The advantage of this graph is that we can allocate frequency resources by simply judging the relationship between the serving clusters in the graph without complicated calculations. Numerical simulation results show that the joint binary user-centric clustering and location-based frequency allocation scheme is superior to the benchmark solutions in terms of the coverage probability.