A Combined Anomaly and Trend Detection System for Industrial Robot Gear Condition Monitoring
Round 1
Reviewer 1 Report
The manuscript presents a methodology selection for condition monitoring of industrial robot gear. However, from the title, it seems that the Authors are offering their own method for condition monitoring. It seems to me that the manuscript is prepared more as a review paper with the goal to pre-select the condition monitoring method.
First of all, the writing style of the manuscript must be changed. With first-person writing style, it seems that the Authors solving some problems that exist in their lab or just giving their opinion on a specific subject (E.g. line 164 //.. which of the models is most suitable for our application... //).
There is no straight research goal defined in the article.
The abstract does not match the rest of the article.
Results are poorly described. Figures 10 and 11 (which are one of the main investigations of the research work) are hard to understand and must be presented in a better manner. Moreover, a more deep analysis of the figures should be presented.
It is very hard to find novelty in that manuscript in its current form.
The straight of the manuscript is a State-of-the-Art section and methods description.
Author Response
Please find attached a detailed letter to your review.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comment 1:
In the abstract, the authors never mentioned gear, but in the problem discussed is about gear, see both in the i Trudy room and title. Also, in the abstract, the Cox-Stuart test is mentioned with any proper introduction. This test is less known to the reader.
Comment 2:
Although it is robots, the essence is a gear system. The problem is fault diagnosis and the determination of the remaining life of the gear system. The authors used the term “trend”. Please explain the difference between the two.
Comment 3:
The generic model of such a problem is (see “A Socially Inspired Framework for Human State Inference Using Expert Opinion Integration”):
- Identification or selection of features,
- Algorithm to build the relationship between the features and target.
The contribution can be either of the above two. Please elaborate on their contribution on the above two aspects.
Comment 4:
Science this journal is applied science, the authors may need to explain their contributions on general inference theory and machine learning.
Author Response
Please find attached a detailed letter to your review.
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors compared the existing machine learning-based method for anomaly and trend detection and proposed a solution for implementing them into industrial robotics. Nevertheless, in the provided manuscript, I noticed few issues:
- First of all, I missed motivation why such study is essential and why it is focused only on industrial robots. The manuscript feels that the authors proposed a good method for abnormalities and trend detection in mechanical components, but the practical implementation is rather weak.
- Authors focus on machine learning-based methods, comparison of those methods with classical techniques used in machine diagnostics missing. Why machine learning fits better? (there are a majority of reliable and simple techniques used in practice).
- An explanation of practical implementation is missing. The whole algorithm of data collection, processing and prediction must be provided in detail. It is unclear what kind of data is used as input or what additional equipment is required to collect it. How proposed method deals with the fact that controllers of industrial robots are typically closed, and the user cannot modify their internal control algorithms?
- Conclusions are not supported by the results, guidelines for further research are unclear.
Author Response
Please find attached a detailed letter to your review.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Dear Authors,
thank you for providing a revised version of the manuscript.
- Next time please highlight the changes that you are making in the manuscript during the resubmission. I understand that with Latex it might be challenging, but you helping reviewers to analyze changes you have done.
- Definition of the research goals and changing writing style helped a lot to deeply understand the value of your paper.
- I do not agree with the way how you present your data. Dot's that you use in Figures 7 and 8 are two small and it is hard to catch the trend. If you want to make those figures understandable for the audience I suggest changing the way you present data and to visualize trendlines.
- Figures 10 and 11 still require your attention. Try to use bigger markers and different shapes of the markers, or something else... at the moment those figures is the weakest part of the paper.
- Figure 14 does not make any sense in that format. I suggest formatting it as well.
Please take the criticism as a suggestion to improve readability and attract more readers to your work. If you publish any figure in your research work it must be understandable even without reading full manuscript.
Author Response
Please find attached a detailed letter to your comments.
Author Response File: Author Response.pdf
Reviewer 2 Report
I am satisfied with the revision along with the rebuttal.
Author Response
Thank you for accepting my changes.
Reviewer 3 Report
In general I am happy with improvements. Only format changes in figures 5, 7, 8-14 required, fonts must be increased. At the moment figures a hardly readable without zooming.
Author Response
Dear reviewer 3,
thank you for your remarks. We increased the font size of the mentioned Figures.
Best regards
Corbinian Nentwich
