.png){kind=icon}
.png){kind=icon}
Return
A Cross-Sensor Interactive Fusion Domain Adaptation Network for Fault Diagnosis in Wheeled Robots Under Nonstationary Motion
.png){kind=icon}
PRE
AI
J
X
T
Y
X
Z
N
X
Y
DOI：10.1109/tim.2026.3716459.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Wheeled robots commonly operate under varying operating conditions and nonstationary motion states, where reliable fault diagnosis is essential for operational safety and long-term service reliability. However, motion-state variations and operating-condition shifts may change intersensor relationships and aggravate cross-domain distribution discrepancies, making fault representations less compact and less class-consistent across different operating scenarios. These factors pose significant challenges to reliable fault diagnosis in practical wheeled robot applications. To address these challenges, a cross-sensor interactive fusion domain adaptation network (CFusion-DAN) is proposed for wheeled robot fault diagnosis under nonstationary motion and varying operating conditions. First, a hierarchical cross-channel fusion strategy (HCFS) is proposed to jointly capture intrasensor temporal dependencies and intersensor coupling, enabling more effective fault feature representation. To further strengthen temporal modeling, a cross-scale interaction module is designed to improve the temporal consistency and hierarchical expressiveness of the fused features. Second, an information-regularized conditional joint alignment (ICJA) mechanism is formulated to suppress intraclass variations induced by operating conditions during marginal and conditional distribution alignment, thereby enhancing classwise compactness and cross-domain discriminability. Extensive experimental evaluations on a real-world wheeled robot platform demonstrate that CFusion-DAN outperforms state-of-the-art approaches across diverse operating scenarios.
Keywords:
Fault diagnosis
multisensor fusion
nonstationary motion
varying operating conditions
wheeled robot
Journal
IF:
5.9
Papers:
1.9W
Citations:
5.8W
Organization
Cited Papers