摘要详情

ID / 提交时间

42 / 2025-05-13 10:58:04

标题

SAGECN: A Spectral-and-GraphSAGE Enhanced Convolutional Network for Robotic Arm Fault Diagnosis

关键字

Motor fault diagnosis,Industrial robotics,Path graph,Graph neural network,Time-series modeling

主题及专题

4、System Monitoring, Fault Diagnosis and Prediction

状态

全文待审

作者

Weixia Hao / Beihang University (Beijing University of Aeronautics and Astronautics)

Haowei Wang / Beihang University (Beijing University of Aeronautics and Astronautics)

Danyang Han / Beihang University (Beijing University of Aeronautics and Astronautics)

Shuohai Sang / Beihang University (Beijing University of Aeronautics and Astronautics)

摘要

Modern industrial robotic systems often operate under dynamic working conditions. Motors in the robotic arm are prone to faults and lead to system instability and performance degradation. Existing fault diagnosis approaches typically rely on static signal features or fixed thresholds, limiting their ability to capture temporal dynamics and structural dependencies. These methods treat sensor channels independently and lack mechanisms to model the temporal evolution and inter-signal correlations, which are critical for detecting subtle and localized fault patterns. This paper proposes a Spectral-and-GraphSAGE Enhanced Convolutional Network (SAGECN) that combines a Graph Convolutional Network and Graph Sample and Aggregate for robotic motor fault diagnosis. Firstly, sensorial time series data is converted into a path graph using a sliding window mechanism to preserve temporal continuity. Then, a multi-stage GNN encoder is constructed by stacking Graph Convolutional Networks (GCN) and Graph Sample and Aggregate (GraphSAGE) layers, enabling the extraction of both global and local structural features. Finally, top-k pooling is applied for hierarchical node selection and graph-level representation learning. Experimental results on a public fault-injected motor dataset demonstrate that the proposed model achieves superior performance in classification accuracy and generalization compared with traditional graph-based methods.