1639 papers
This paper proposes a dual-stream multiscale attention network (MS-BiGRU-Attn) that processes axial and normal magnetic flux leakage signals independently to accurately invert the width of composite pipeline defects under nonlinear magnetic interference, achieving superior performance on both out-of-distribution simulations and experimental validation compared to baseline models.
This paper proposes a mode-selective transmission method for guided waves that utilizes data processing techniques, including group-velocity and phase synchronization along with wave-packet excitation, to selectively enhance specific modes using conventional piezoelectric transducers without requiring hardware or procedural modifications.
This paper presents a reactive 3D motion planning approach for a Franka Emika Panda manipulator that clusters overlapping inflated obstacles into star-shaped proxies to enable dynamical system-based control, demonstrating improved goal attainment in complex scenarios while highlighting trade-offs such as increased path length and potential navigable corridor closures.
This study integrates Fire Dynamics Simulator (FDS) numerical simulations, full-scale smoke tests, and a Long Short-Term Memory (LSTM) deep learning model to analyze smoke dispersion in metro platform fires and develop a highly accurate early warning system that achieves over 98% accuracy in predicting critical visibility drops.
This study demonstrates that a fine-tuned ResNet-18 convolutional neural network can effectively automate the distinction between genuine process outliers and measurement misdetections in optical coherence tomography (OCT) data, significantly accelerating pharmaceutical coating development by reducing manual review time and improving data reliability.
Reliableformer is a device-aware, physically-consistent digital twin framework that maps Vision Transformers onto memristor crossbars to simulate non-idealities, evaluate trade-offs, and apply circuit- and software-level mitigations, thereby enabling reliable and sustainable analog AI acceleration.
This paper presents a reproducible Python-based framework for automatically assessing the primary frequency response of synchronous hydrogenerators against Ecuador's ARCONEL-001/24 grid code, demonstrating its ability to detect controller faults and calculation inconsistencies through a case study of 100-MVA units with implications for mining-captive power systems.
This paper proposes an adjacent-four-vector synthesis-based PWM strategy that enhances common-mode voltage suppression and output harmonic performance by optimizing switching sequences and leveraging increased control degrees of freedom compared to traditional nearest-three-vector methods.
This paper proposes a simple and easily applicable reference model for the Internet of Things that focuses on communication technology to address the complexity of existing models while supporting the secure and ethical processing of vast data in the merging physical and digital worlds.
This study investigates shear strain localization in Zn-Mg wire during rotational dieless drawing by combining numerical FEM modeling with experimental validation, revealing that shear strain localizes similarly to axial strain at critical levels and high temperatures but manifests as increased surface roughness rather than significant geometric changes.