Industrial Robot Fault Detection Based on Statistical Control Chart

Jaber, Alaa Abdulhady; Bicker, Robert

doi:10.3844/ajeassp.2016.251.263

Cited by 12 publications

(5 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They detect data points that deviate from the distribution of the historical data. These methods used for robots include Statistical Control Charts (SCCs) [9], Principal Component Analysis (PCA) based method [10], Partial Least Squares (PLS) based approach [11] and so on. However, most of these methods require that all the data have to be accumulated before faults can be detected, which make them unsuitable for real-time anomaly detection.…”

Section: B Approachesmentioning

confidence: 99%

“…The main task of CBM is to perform real-time anomaly detection from the gathered time series data, which takes note of indicative fault data that do not conform to some explicit laws or historical patterns. There are two general techniques usually used for anomaly detection: model-based techniques [5]- [8] and data-driven techniques [2], [9]- [20]. Model-based techniques have been developed due to their good performance in predicting specific faults of the robots.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

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

et al. 2020

View full text Add to dashboard Cite

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 shut down. Model-based methods are typically used in anomaly detection for robots, yet explicit domain knowledge and accurate mathematical models are required. Data-driven techniques can overcome these limitations. However, a major difficulty for them is the lack of sufficient fault data of industrial robots. Besides, the used technique for anomaly detection of robots should be required to not only capture the temporal dependency in collected time series data, but also the inter-correlations between different metrics. In this paper, we introduce an unsupervised anomaly detection for industrial robots, sliding-window convolutional variational autoencoder (SWCVAE), which can realize real-time anomaly detection spatially and temporally by coping with multivariate time series data. This method has been verified by a KUKA KR6R 900SIXX industrial robot, and the results prove that the proposed model can successfully detect anomaly in the robot. Thus, this work presents a promising tool for condition-based maintenance of industrial robots.INDEX TERMS Anomaly detection, industrial robots, sliding window, variational autoencoder, convolutional neural network.

show abstract

Section: B Approachesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

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

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Examples on the utilization of control charts for industrially used robots in manufacturing environments are given in [9], [10]. Furthermore the authors in [11] use control charts for fault detection in manufacturing robots. They monitor the wear and tear of a robot arm by logging the vibrations in a control chart in order to monitor faults at early stages.…”

Section: Related Workmentioning

confidence: 99%

Metric Based Dynamic Control Charts for Edge Anomaly Detection in Factory Logistics

Eschemann

Borchers²,

Lisiecki

et al. 2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

The optimization of transport logistics in production environments is a holistic task for the factory of the future. Autonomous guided vehicles that perform transport jobs in factories are facing this challenge and have to detect, react and prepare to unforeseen changes and anomalies in the production system. Due to data protection concerns, details like production plans are often not available for an external transportation system. Hence the anomaly detection has to be based on self-collected and observed data of the transport system like occurred transport needs or the evolution of internal metrics. In this paper we infused a production system with manufacturing process anomalies and demonstrate a detection based on the observation of transport needs to overcome the gab caused by restricted information. For that detection we extended classic control charts to work with expected values based on learned dynamic production characteristics. The system sets a tolerance field as narrow as possible around dynamically determined values, resulting in an average precision of 95% for detection unusual number of transport jobs.

show abstract

“…During the four years of the first mission, the ship scanned 150,000 stars, searching for the small falls in light caused by planets that pass in front of them. In 2014, Kepler switched to the second mission called K2, where he still hunts planets but also makes a variety of observations (Wang and Yagi, 2016;Obaiys et al, 2016;Ahmed et al, 2016;Jauhari et al, 2016;Syahrullah and Sinaga, 2016;Shanmugam, 2016;Jaber and Bicker, 2016;Moubarek and Gharsallah, 2016;Amani, 2016;Shruti, 2016;Pérez-de León et al, 2016;Mohseni and Tsavdaridis, 2016;Abu-Lebdeh et al, 2016;Serebrennikov et al, 2016;Budak et al, 2016;Augustine et al, 2016;Jarahi and Seifilaleh, 2016;Nabilou, 2016;You et al, 2016;Zurfi and Zhang, 2016;Rama et al, 2016;Sallami et al, 2016;Huang et al, 2016;Ali et al, 2016;Kamble and Kumar, 2016;Saikia and Karak, 2016;Zeferino et al, 2016;Pravettoni et al, 2016;Bedon and Amadio, 2016;Chen and Xu, 2016;Mavukkandy et al, 2016;Yeargin et al, 2016;Madani and Dababneh, 2016;Alhasanat et al, 2016;Elliott et al, 2016;Suarez et al, 2016;Kuli et al, 2016;…”

Section: The Incredible Quantitymentioning

confidence: 99%

The TESS Satellite will Search for Planets in the Vicinity of our Solar System

Petrescu¹

2019

Journal of Aircraft and Spacecraft Technology

View full text Add to dashboard Cite

The US Space Agency has launched from Cape Canaveral, Florida, the TESS satellite, which will search for planets near our solar system and study them in detail. "TESS will significantly increase the number of planets studied," said George Ricker of the Massachusetts Institute of Technology, MIT, before launching. "It will double the number of planets we've seen and detected with the Kepler telescope," he added. The Kepler Space Observatory has discovered 2,650, about 70% of the total worlds outside the solar system that is known to mankind. TESS was launched with a Falcon 9 rocket, which returned to Earth after sending the satellite to the Earth's orbit. The telescope used by NASA to "hunt" the planets led to the discovery of two heavenly bodies that seem to be ideal places for life. The two planets have the ideal dimensions and are at the right distance from their star. William Borucki, the Kepler telescope scientist, has announced that these two planets are the best candidates identified so far in the race to discover planets that host life. The findings were published in the journal Science. Experts announce that these two planets are an important milestone in the race for identifying extraterrestrial life. In the past, when astronomers identified exoplanets (planets outside our solar system) they did not meet all the criteria necessary to accommodate life forms. Many planets are not in the habitable area-at a distance from their star so they are not too hot or too cold, but ideal for the existence of liquid water. Until now, all the planets identified inhabitable areas were too large. Most likely, these planets were huge gas balls, like Neptune, not suitable for life. Instead, the Earth-sized planets discovered so far have been at a distance unsuitable for their stars. The two planets now identified by the researchers, Kepler-62-e and Kepler-62-f, meets all the necessary conditions for hosting life. The two planets are twin-they orbit the same star and are next to each other, less than Earth and Mars. Planets are a bit wider than Terra, but they are not too big. Kepler-62-e is a warm Hawaiian planet and Kepler-62-f is cool like Alaska, Borucki explained. Researcher David Charbonneau of Harvard University, co-author of this study, says, "It's the first planet I really think of," he said, "but Kepler-62-f may be able to host life forms." The researcher added that "we have overcome an important barrier. So, why not have life forms?" The planets are 1,200 light-years away from Terra. The star that it orbits dates back 7 billion years ago, being 2.5 billion years older than our Sun.

show abstract

Industrial Robot Fault Detection Based on Statistical Control Chart

Cited by 12 publications

References 16 publications

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

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

Metric Based Dynamic Control Charts for Edge Anomaly Detection in Factory Logistics

The TESS Satellite will Search for Planets in the Vicinity of our Solar System

Contact Info

Product

Resources

About