报告详情
  1. 活动首页
  2. 报告清单
  3. 报告详情
A Multi-stage Multi-scale Feature Fusion Network Based on KAN for Industrial Robot Fault Diagnosis
编号：123 访问权限：仅限参会人 更新：2025-11-10 15:45:51 浏览：65次 张贴报告

报告开始：暂无开始时间（Asia/Shanghai）

报告时间：暂无持续时间

所在会场：[暂无会议] [暂无会议段]

暂无文件

摘要
With the rapid advancement of China's manufacturing sector, industrial robots have assumed a pivotal role within the industry. The operational status of these robots directly impacts a company's production efficiency. However, the complex working environments of industrial robots pose significant challenges for existing deep learning models in extracting critical information from signals. To address this issue, this paper proposes a multi-stage, multi-scale feature fusion network based on KAN. Firstly, one-dimensional signals are transformed into wavelet time-frequency maps, and a multi-scale information flow is designed based on LSConv. Secondly, to fully extract critical features from the multi-scale information flow, an attention-based LSConv feature enhancement layer is designed. Subsequently, to leverage features from multiple network stages, a multi-scale feature enhancement layer is implemented. Finally, KAN is employed as a classifier to further enhance the network's diagnostic capabilities. Experimental results on the SDUST dataset demonstrate that the proposed method outperforms existing fault diagnosis approaches.
关键词
Kolmogorov-Arnold network(KAN),Fault Diagnosis,Attention mechanism,Multi-stage Feature,Multi-scale feature
报告人
Junjie He
Doctoral student Southeast University

稿件作者
Junjie He Southeast University
Lingfei Mo Southeast University
发表评论
验证码 看不清楚，更换一张
全部评论
重要日期

  • 会议日期

    11月21日

    2025

    至

    11月23日

    2025

  • 10月20日 2025

    初稿截稿日期

  • 11月23日 2025

    注册截止日期

主办单位
IEEE Instrumentation and Measurement Society
South China University of Technology
承办单位
South China University of Technology
联系方式
登录查看完整联系方式
历届会议
移动端
在手机上打开
小程序
打开微信小程序
客服
扫码或点此咨询