Multiple Input Multiple Output (MIMO) technology significantly enhances signal transmission rates and system reliability through multi-antenna systems. To improve spectral efficiency and anti-interference capabilities, spatial modulation technology, as an extension of MIMO, has been proposed and widely applied. Generalized Spatial Modulation (GSM) further integrates multiple modulation schemes, enhancing the system's performance. Polar codes, as an efficient error correction code, leverage channel polarization to transform physical channels into virtual channels with varying levels of reliability, thus effectively improving the performance of MIMO and spatial modulation systems. This paper presents a decoding scheme for multi-user polar codes, aimed at optimizing the decoding process in the uplink Polar Coded-Generalized Spatial Modulation (PC-GSM) system. By combining the channel polarization characteristics of polar codes with the advantages of GSM, the scheme improves decoding algorithms, enhancing the reliability and data transmission rate of multi-user systems. Simulation results show that the proposed decoding scheme significantly boosts system performance, providing a novel solution for the integration of multi-user polar codes and spatial modulation technology.

The Internet of Things (IoT), as one core area of 6G development, plays a crucial role in driving network architecture changes and supporting core application scenarios. However, IoT systems suffer from energy imbalances and short network lifecycles, which severely restrict the improvement of data collection efficiency. With the rise of Unmanned Aerial Vehicle (UAV) technology, its high maneuverability can effectively construct Line of Sight (LOS) communication links, thereby improving communication speed. This has great application value in data collection of IoT systems and can solve the problem of low data collection efficiency caused by the short lifecycle of IoT networks. To this end, UAVs are used to collect data from ground IoT devices and build a data collection and transmission link for air-to-ground collaboration. An intelligent data collection method based on Deep Reinforcement Learning (DRL) is proposed. In addition, a predictive neural network is designed to further improve data collection efficiency by predicting network data at the Base Station (BS) side, thereby achieving the goal of reducing IoT device energy consumption and extending network lifespan. Simulation results show that the proposed data collection algorithm has good performance advantages in terms of device energy consumption and energy balance, and is superior to traditional data collection algorithms. At the same time, the proposed data collection network architecture can extend the network lifespan by 1.2 times when the predicted data accounts for 12.5%. In addition, simulations have shown that the designed predictive neural network outperforms other compared networks in terms of Mean Squared Error (MSE) and Mean Absolute Error (MAE) metrics.

Focusing on the channel estimation accuracy degradation caused by the beam splitting effect in near-field wideband Extremely-Large Scale Multiple Input Multiple Output (XL-MIMO) systems, this paper proposes a Bi-Directional Integrated Multi-Subcarrier Augmented Bilinear Pattern Detection (BDI-MSABPD) algorithm. Built upon the polar-domain sparse representation framework, the proposed method addresses both the sparse support set misalignment and parameter estimation bias induced by beam splitting through a dual mechanism combining explicit polar-domain resolution enhancement and implicit multi-subcarrier joint optimization. Simulation results demonstrate that the BDI-MSABPD achieves an average reduction of 2 dB in Normalized Mean Squared Error (NMSE) compared with conventional Bilinear Pattern Detection (BPD) algorithm.

Skin cancer and melanocytic nevus share numerous similarities, which can result in a misdiagnosis by dermatologists. To improve the screening accuracy of early skin cancer patients, the Gamma Transform Block (GMTB) based on Gamma Transform (GT) and Wavelet Convolution Block (WTCB) based on Wavelet Transform (WT) are proposed. Furthermore, the Space-Frequency Transform Network (SFTNet) for capturing fine-grained features of skin cancer is innovatively proposed based on the Detection Transformer(DETR) architecture. SFTNet-based skin cancer screening system can effectively improve disease detection accuracy because it enhances the sample image at different channels and reduces over-fitting effect during the model training process. Simulation results on HAM10000 dataset show that the accuracy of this system can reach 85.5%, which underscores the significant clinical value of our approach in skin cancer assisted diagnosis.

Chaos-based communication technology has emerged as a research hotspot in recent years due to its superior resistance to multipath fading and robust security features. Differential Chaos Shift Keying (DCSK), as a non-coherent digital modulation scheme, has attracted widespread attention. However, in practical communication scenarios, the increasing demand for reliable data transmission has revealed the limitations of traditional DCSK systems, such as low transmission rates and high Bit Error Ratios (BER), highlighting the urgent need to enhance system reliability. Considering the significant advantages of polar codes, including low complexity and near-capacity performance, this paper delves into the integration of polar coding algorithms with chaos modulation techniques based on channel polarization principles, aiming to further improve the reliability of chaos-based communication systems. Experiment results show that the proposed solution significantly improves the reliability of chaos-based communication systems and keeps its feature of low complexity.

Reconfigurable Intelligent Surface (RIS) and Rate Splitting Multiple Access (RSMA) technologies are two emerging communication techniques with broad prospects in future wireless systems. Active RIS (ARIS) has the advantage of overcoming the effects of multiplicative fading compared to Passive RIS (PRIS). The paper addresses the outage performance of ARIS-assisted RSMA systems under random deployment. Under Nakagami-m small-scale fading model in downlink, multiple users are randomly distributed within a semicircular region around ARIS and sorted by the distances between the users and ARIS. Taking the optimal phase shift for the furthest user as the entry point, a moment matching approach is employed to derive the shape and scale parameters associated with the cascaded channel power characteristic for each RSMA user. We derive a closed-form expression for the outage performance of the users, analyze the effect of the power splitting coefficients on the Outage Probability (OP), and obtain the diversity order and an expression for the approximate OP. Simulation results show that the OP of a non-perfect phase-shifted user under ARIS-assistance is re duced by 58% compared to the passive RIS-assisted system under a power budget of 50 dBm and a moderate amount of RIS elements. In addition, the power consumption of the user with the optimal phase shift is reduced by 53 dBm.

To address the challenges of data scarcity, stylistic diversity, and complex textures in art image classification, a novel self-supervised learning framework is proposed——Frequency-Masked Contrast (F-MaCo). Built upon a dual-branch contrastive learning paradigm, F-MaCo leverages a two-dimensional Discrete Wavelet Transform (DWT) to project images into the frequency domain, enabling dynamic frequency-domain masking augmentation. Additionally, a perceptual loss-driven weighting mechanism is introduced to effectively capture the multi-scale features and rich textures information of art images. Experimental results demonstrate that F-MaCo achieves state-of-the-art performance on four art image datasets—MAMe, Kaokore, Artbench10, and ArtDL—with Top1 accuracies of 73.72%, 77.38%, 58.38%, and 68.31%, respectively, validating its effectiveness and robustness in art image representation learning.

Existing key point detection algorithms tend to suffer from reduced detection precision, missed detections, or misaligned key points in scenarios with varying lighting conditions and dense crowds with overlapping figures. To address this issue, an improved LBW-YOLOv8n-Pose algorithm for multi-person pose estimation in complex environments is proposed based on YOLOv8n-Pose. By introducing the Large Separable Kernel Attention (LSKA) in the Spatial Pyramid Pooling-Fast (SPPF) layer of the feature extraction backbone network, the algorithm enhances the image feature representation and perception capabilities. A weighted Bidirectional Feature Pyramid Network (BiFPN) is incorporated in the neck network for reconstruction to improve the multi-scale feature fusion effect. Additionally, an improved Wise-IoU loss function is adopted to accelerate the model's convergence speed and enhance its robustness in complex scenarios. Experimental results show that the improved model achieves precision, recall, and average detection precision of 85.7%, 76.8%, and 81.7% respectively on the MS-COCO2017 human key point dataset, representing significant improvements over the original model. Moreover, it can more accurately and effectively detect key point information of multiple people in complex situations.

In contemporary society, Global Navigation Satellite System (GNSS) has become an essential tool for daily travel, significantly improving travel efficiency. However, in environments with weak signals, such as indoors or tunnels, GNSS systems often experience signal loss due to insufficient signal strength, leading to positioning failure and inability to provide accurate navigation services. To address this challenge, this paper proposes a high-precision positioning solution based on an improved Extended Kalman Filter (EKF). This solution integrates Ultra-Wideband (UWB) least squares method, KF, and EKF technologies, and introduces a Multi-Innovation EKF (MIEKF) algorithm. By utilizing multi-time observation data and a forgetting factor mechanism, the solution effectively reduces positioning errors and enhances positioning accuracy. Experimental results show that the root mean square error of this solution can be reduced to 0.179 m, verifying its high-precision positioning capability in weak signal environments and providing reliable technical support for precise navigation in complex scenarios.

Compared with a different-color backgrounds, recognizing and detecting cucumber fruits under uniform-color backgrounds remains a key challenge due to limited distinguishing features and increased susceptibility to occlusion and background interference. To address this, we propose YOLO-ACG, a detection network based on YOLOv11n. An Adaptive Dynamic Downsample (A-Down) module is introduced, combining deformable convolution and channel attention to achieve adaptive cross-scale feature sampling. A Ghost_HGNetV2 architecture is designed, where the High-resolution Group Stem (HGStem) reduces input channels to extract efficient intrinsic features, and the Ghost_HGBlock applies knowledge distillation to enhance feature representation. A Context and Spatial Feature Calibration Network (CSFCN) network structure is introduced, which includes Context Feature Calibration (CFC) and Spatial Feature Calibration (SFC). The CFC module aggregates context information relevant to each pixel, while the SFC module leverages calibrated spatial features to ensure accurate understanding of spatial layout the image. Together, they enable the network to more precisely distinguish cucumber fruits from backgrounds with similar colors. Experimental results show that the improved model achieves 4.64 percentage points increase in precision, recall by 5.07 percentage points, F1 by 4.89 percentage points, and mAP by 4.48 percentage points. Ablation and comparative experiments confirm that YOLO-ACG significantly reduces false positives and missed detections, offering effective technical support for cucumber fruits recognition in complex, uniform-color environments.