2020
DOI: 10.1109/access.2020.2977892
|View full text |Cite
|
Sign up to set email alerts
|

Unsupervised Anomaly Detection of Industrial Robots Using Sliding-Window Convolutional Variational Autoencoder

Abstract: With growing dependence of industrial robots, a failure of an industrial robot may interrupt current operation or even overall manufacturing workflows in the entire production line, which can cause significant economic losses. Hence, it is very essential to maintain industrial robots to ensure high-level performance. It is widely desired to have a real-time technique to constantly monitor robots by collecting time series data from robots, which can automatically detect incipient failures before robots totally … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

2
62
0

Year Published

2020
2020
2024
2024

Publication Types

Select...
5
3
2

Relationship

1
9

Authors

Journals

citations
Cited by 132 publications
(64 citation statements)
references
References 23 publications
2
62
0
Order By: Relevance
“…This method divides sequences into many overlapping subsequences and then calculates the correlations among the subsequence data in a window by a given algorithm. Chen et al [21] proposed a sliding window convolutional differential autoencoder that can detect the anomalies of multivariate time-series in time and space. When processing multidimensional time-series, we also need to consider the characteristics of the data itself in addition to the correlation between series.…”
Section: Related Workmentioning
confidence: 99%
“…This method divides sequences into many overlapping subsequences and then calculates the correlations among the subsequence data in a window by a given algorithm. Chen et al [21] proposed a sliding window convolutional differential autoencoder that can detect the anomalies of multivariate time-series in time and space. When processing multidimensional time-series, we also need to consider the characteristics of the data itself in addition to the correlation between series.…”
Section: Related Workmentioning
confidence: 99%
“…By coping with electrocardiography (ECG) data, an ensemble of autoencoders joined by the sparse connections of recurrent neural networks (RNNs) was developed [36]. CAE and LSTM AE that utilized a two-stage sliding window strategy for time series data preprocessing showed successful results for IoT data [37], while a sliding-window convolutional VAE was adopted to capture spatio-temporal patterns and detect anomalies in real-time for industrial robots [38]. However, most previous works merely average the reconstruction error and employ it as an anomaly score.…”
Section: Related Workmentioning
confidence: 99%
“…Fault detection in an industrial environment has always been a challenging task [20], [21], [22], [23], [24]. Due to the issues related to interoperability and communication between different devices, collecting a large dataset in such an environment is not an easy task.…”
Section: State Of the Artmentioning
confidence: 99%