Diagnosis of mechanical faults in manufacturing systems is critical for ensuring safety and saving costs. With the development data transmission sensor technologies, measuring can acquire massive amounts multisensor data. Although deep learning (DL) provides an end-to-end way to address drawbacks traditional methods, it necessary do research on intelligent fault diagnosis method based In this article, a novel fusion (MSF) convolutional neural network (CNN) explored. First,...

This robot uses STM32F103RCT6 as the control core and relies on ducted fan power primary ball-picking mechanism. The ball is collected using inertia of moves forward suction from fan, allowing to enter collection box above robot. completes process when it across all positions within field picks up balls in its path. It offers two modes: manual fully automatic. integrates multiple functions, including remote control, Bluetooth, tracking, distance sensing obstacle avoidance, power, solar...

Abstract Melt pool modeling is critical for model-based uncertainty quantification (UQ) and quality control in metallic additive manufacturing (AM). Finite element (FE) simulation thermal metal AM, however, tedious time-consuming. This paper presents a multifidelity point-cloud neural network method (MF-PointNN) surrogate of melt based on FE data. It merges the feature representations low-fidelity (LF) analytical model high-fidelity (HF) data through theory transfer learning (TL). A basic...

Multisensory systems play a critical role in prognostics and health management (PHM), utilise the information from multi-device synchronous measurements for fault diagnosis predictive maintenance. But it is not suitable specific with limited bandwidth energy reservoirs since increased sophistication of measurement devices requires more computation power resources. This research explores data-driven analytical framework multisensory system analysis design PHM. The proposed provides optimal...

Abstract Intelligent fault diagnosis (IFD) techniques commonly use vibration-based measurements to perform health monitoring of critical rotating components in industrial systems. However, these approaches may be limited cost-sensitive applications, because the installation vibration sensors is inconvenient and are expensive. Considering difficulties IFD using only current-related information from motor current signal (MCS), this paper proposes a three-dimensional hybrid-fusion neural...

Abstract The small changes in process parameters have significant influences on the stability of laser powder bed fusion (LPBF). Therefore, monitoring is particularly important. This paper proposed a machine learning (ML)-based multi-sensor approach to monitor LPBF processing state by combining photodiode, acoustic, and visual signals. In order extract motion features melt pool more accurately describe its transient changes, an ellipse adjustment algorithm segment images, eliminating...

Single track is the basis for melt pool modeling and physics work in laser powder bed fusion (LPBF). The melting state of a single closely related to defects such as porosity, lack fusion, balling, which have significant impact on mechanical properties an LPBF-created part. To ensure reliability part quality repeatability, process monitoring feedback control are emerging improve states, becoming hot topic both industrial academic communities. In this research, simple low-cost off-axial...

Abstract With the development of deep learning and information technologies, intelligent welding systems have been further developed, which achieve satisfactory identification defective welds. However, lack labeled samples complex working conditions can hinder improvement models. This paper explores a novel method based on metric-based meta-learning for classification defects with cross-domain few-shot (CDFS) problems. First, an embedding module using convolutional neural network (CNN) is...

(1. 广西大学机械工程学院 南宁 530004； 2. 广西大学广西制造系统与先进制造技术重点实验室 530004)

Multi-modality fusion is proven an effective method for 3d perception autonomous driving. However, most current multi-modality pipelines LiDAR semantic segmentation have complicated mechanisms. Point painting a quite straight forward which directly bind points with visual information. Unfortunately, previous point like methods suffer from projection error between camera and LiDAR. In our experiments, we find that this the devil in painting. As result of that, propose depth aware mechanism,...

The aim of this study is to address the limitations in reconstructing electrical activity heart from body surface electrocardiogram, which an ill-posed inverse problem. Current methods often assume values commonly used literature absence

We introduce a novel MV-DETR pipeline which is effective while efficient transformer based detection method. Given input RGBD data, we notice that there are super strong pretraining weights for RGB data less works depth related data. First and foremost , argue geometry texture cues both of vital importance could be encoded separately. Secondly, find visual feature relatively hard to extract compared with in 3d space. Unfortunately, single dataset thousands not enough training an...

Abstract Intelligent diagnostic models with deep learning networks have been widely used for health conditions identification of complex mechanical systems. However, the distributional shift and data scarcity phenomenon inevitably exist in practical real-world scenarios, which further affect generalization representation capability neural networks. Recently, meta-learning has provided an effective way to solve domain shifts extremely limited problems. By introducing non-parametric similarity...

The active protection system (APS), usually installed on the turret of armored vehicles, can significantly improve vehicles' survivability battlefield by launching countermeasure munitions to actively intercept incoming threats. However, uncertainty over launch angle is increased due angular disturbances when off-road vehicle moving rough terrain. Therefore, accurate and comprehensive disturbance prediction essential real-time monitoring angle. In this paper, a deep ensemble learning...