The existing fixed gait lower limb rehabilitation robots perform a predetermined walking trajectory for patients, ignoring their residual muscle strength. To enhance patient participation and safety in training, this paper aims to develop robot with adaptive training capability relying on human-robot interaction force measurement. Firstly, novel system several active passive driven joints is developed, 2 face-to-face mounted cantilever beam sensors are employed measure the forces. Secondly,...

In response to the issues that existing voltage regulation methods cannot finely perceive load data within station area, have low precision, and often use active power leading high electricity costs for users, a novel intelligent method area based on multi-agent reinforcement learning is proposed. Firstly, photovoltaic (PV) output prediction technology hybrid neural network proposed predict previous day. Secondly, strategy generation mechanism integrates reactive control designed, this, an...

Quantum nanostructures offer crucial applications in electronics, photonics, materials, drugs, etc. For accurate design and analysis of simulations the Schrodinger or Schrodinger-like equation are always needed. large nanostructures, these eigenvalue problems can be computationally intensive. One effective solution is a learning method via Proper Orthogonal Decomposition (POD), together with ab initio Galerkin projection equation. POD-Galerkin projects problem onto reduced-order space POD...

Map construction task plays a vital role in providing precise and comprehensive static environmental information essential for autonomous driving systems. Primary sensors include cameras LiDAR, with configurations varying between camera-only, LiDAR-only, or camera-LiDAR fusion, based on cost-performance considerations. While fusion-based methods typically perform best, existing approaches often neglect modality interaction rely simple fusion strategies, which suffer from the problems of...

The fast and stable reconstruction of building interiors from scanned point clouds has recently attracted considerable research interest. However, reconstructing long corridors connected areas across multiple floors emerged as a substantial challenge. This paper presents comprehensive segmentation method for three-dimensional (3D) indoor structure with stories. With this method, the over-segmentation that usually occurs in complex environment is overcome by morphologically eroding floor...

Oxygen is an important energy medium in the steelmaking process. The accurate dynamic prediction of oxygen demand needed to guarantee molten steel quality, improve production rhythm, and promote collaborative optimization energy. In this work, a analysis mechanism industrial big data was undertaken, we found that characteristic factors Basic Furnace (BOF) consumption were different modes, such as duplex dephosphorization, decarbonization, traditional mode. Based on this, dynamic-prediction...

As the foundation for digitalization, building information modeling (BIM) technology has been widely used in field of architecture, engineering, construction, and facility management (AEC/FM). Unmanned aerial vehicle (UAV) oblique photogrammetry laser scanning have become increasingly popular data acquisition techniques surveying buildings providing original BIM modeling. However, geometric topological reconstruction solid walls, which are among most important architectural structures BIM,...

Pre-trained language models (PLMs) have greatly transformed various downstream tasks, yet frequently display social biases from training data, raising fairness concerns. Recent efforts to debias PLMs come with limitations: they either fine-tune the entire parameters in PLMs, which is time-consuming and disregards expressiveness of or ignore reintroducing tasks when applying debiased them. Hence, we propose a two-stage pipeline mitigate both internal contexts while preserving models....

The automatic identification and classification of image-based weld defects is a difficult task due to the complex texture X-ray images defect.Several depth learning methods for automatically identifying welds were proposed tested.In this work, four different convolutional neural networks evaluated compared on 1631 image set.The concavity, undercut, bar defects, circular unfused incomplete penetration in 6 types are classified.Another contribution paper train CNN model "RayNet" dataset from...

Superpixel segmentation is an important image preprocessing technology that can aggregate adjacent pixels with similar characteristics to reduce the complexity of subsequent tasks. However, increasing spatial resolution remote sensing images, maintaining accurate boundary information ground objects during superpixel has become a challenge. We propose feature reconstruction method based on edge detection and adaptive morphological enhance representation for segmentation. Specifically, we...

As the rapid development of earth observation technology and deep learning, building extraction from remotely sensed imagery based on convolutional neural networks (DCNNs) has attracted wide attention in recent years. However, due to heterogeneity shapes sizes complexity surrounding objects, current methods still have challenges boundary accuracy complete extraction. For these purposes, we proposed low-level feature enhancement multi-scale pyramid aggregation network (LFEMAP-Net) that...

Few-shot scene classification methods aim to obtain discriminative ability from few labeled samples and has recently seen substantial advancements. However, the current few-shot learning approaches still suffer overfitting due scarcity of samples. To this end, a semi-supervised method is proposed address issue. Specifically, used increase target domain samples; then we train multiple models using augmented Finally, perform decision fusion results obtained accomplish image task. According...

<title>Abstract</title> Convolutional neural network (CNN) is frequently used in image classification. However, obtaining sufficient labelled data for training difficult because labelling costly. Learning from a limited number of samples creates challenges since the learned model may become overfitted due to biased distribution caused by few and pow learning algorithm. This paper proposed dynamic calibration method shot-learning. First, base new class were normalized using normalization...

With the rapid development of industrial production, defect detection has become increasingly important across various fields. However, traditional methods suffer from issues such as low efficiency, high costs, and lack real-time capabilities. To address these problems, this paper proposes an intelligent system based on cloud-edge synergy. This leverages advantages cloud computing edge to achieve efficient, cost-effective, detection. Specifically, first trains models servers then distributes...

In this paper, we present a novel approach for multiview point cloud registration. Different from previous researches that typically employ global scheme registration, propose to adopt an incremental pipeline progressively align scans into canonical coordinate system. Specifically, drawing inspiration image-based 3D reconstruction, our first builds sparse scan graph with retrieval and geometric verification. Then, perform registration via initialization, next selection Track create continue,...

Cycling has gained global popularity for its health benefits and positive urban impacts. To effectively promote cycling, early studies have extensively investigated the relationship between cycling behaviors environmental factors, especially cyclists' preferences when making route decisions. However, these often struggle to comprehensively describe detailed procedures at a large scale due data limitations, they tend overlook complex nature of preferences. address issues, we propose novel...