The key to improving the performance of graph convolutional networks (GCN) is to fully explore the correlation between neighboring and distant information. Aiming at the over-smoothing problem of GCN, in order to make full use of the relationship among features, graphs and labels, a graph residual generation network based on multi-information aggregation (MIA-GRGN) is proposed. Firstly, aiming at the defects of GCN, we design a deep initial residual graph convolution network (DIRGCN), which connects the initial input through residuals, so that each layer node retains part of the information of the initial features, ensuring the localization of the graph structure and effectively alleviating the problem of over-smoothing. Secondly, we propose a random graph generation method (RGGM) by utilizing graph edge sampling and negative edge sampling, and optimize the supervision loss function of DIRGCN in the form of generation framework. Finally, applying RGGM and DIRGCN as inference modules for modeling hypotheses and obtaining approximate posterior distributions of unknown labels, an optimized loss function is obtained, we construct a multi-information aggregation MIA-GRGN that combines graph structure, node characteristics and label joint distribution. Experiments on benchmark graph classification datasets show that MIA-GRGN achieves better classification results compared with the benchmark models and mainstream models, especially for datasets with less dense edge relationships between nodes.

This paper presents a new enhanced coprime array for direction of arrival (DOA) estimation. Coprime arrays are capable of estimating the DOA using coprime properties and outperforming uniform linear arrays. However, the associated algorithms are not directly applicable for estimating the DOA of coherent sources. To overcome this limitation, we propose an enhanced coprime array in this paper. By increasing the number of array sensors in the coprime array, it is feasible to enlarge the aperture of the array and these additional array sensors can be utilized to achieve spatial smoothing, thus enabling estimation of the DOA for coherent sources. Additionally, applying the spatial smoothing technique to the signal subspace, instead of the conventional spatial smoothing method, can further improve the ability to reduce noise interference and enhance the overall estimation result. Finally, DOA estimation is accomplished using the MUSIC algorithm. The simulation results demonstrate improved performance compared to traditional algorithms, confirming its feasibility.

In response to the challenges posed by existing integrated localization schemes in Global Navigation Satellite System (GNSS)-denied scenarios, such as the reliance on prior high-precision maps and inadequate flexibility of localization paths, this paper presents an integrated localization method employing triggered magnetic correction. This method leverages the Inertial Measurement Unit (IMU) to assist both the camera and the Light Detection and Ranging (LiDAR) systems, functioning as a relative positioning subsystem to compensate for precision loss in the blind zones of magnetic landmark signals. By integrating Magnetic Guidance Sensor (MGS) and Radio Frequency Identification (RFID) tags, triggered magnetic correction has been successfully implemented, providing high-precision global information constraints. An enhanced multimodal data fusion algorithm has been designed based on a dual-level filtering architecture, integrating relative and absolute positioning. The experimental results of the free-path scenario demonstrate that the average positional error for the full path is 2.94 cm, and the average angular error is 0.63°. Compared to results without magnetic correction, the average positional error and average angular error decrease by 68.53% and 64.99%, respectively. These results confirm the feasibility and effectiveness of the proposed method.

Profound heterogeneity in Autism Spectrum Disorder (ASD) complicates diagnosis and the development of effective treatments. In healthcare systems with limited specialist resources, the need for rapid and accessible screening tools is particularly urgent. In the present study, we developed an objective, scalable pipeline that pairs a simple two-minute video recording of a child’s naturalistic behavior with a deep AI model for autism detection, representing the first such rapid, scalable framework. By analyzing a rich spectrum of children’s responses to social stimuli, such as subtle behavioral patterns, gaze dynamics, facial morphology, and dynamic facial complexity, the deep AI model provides powerful support for clinical workflows, demonstrating high accuracy in identifying ASD risk across diverse internal and external test cohorts, irrespective of sex, age, or cognitive function. Furthermore, a series of comprehensive analyses confirmed the model’s clinical relevance and revealed its capacity to objectively stratify ASD heterogeneity into neurobiologically distinct subtypes. This work establishes a highly efficient and objective framework for large-scale screening, providing a data-driven foundation to stratify heterogeneity and paving the way for the future development of targeted interventions.