Detection of Operator Fatigue in the Main Control Room of a Nuclear Power Plant Based on Eye Blink Rate, PERCLOS and Mouse Velocity

Round 1

Reviewer 1 Report

• The term PERCLOS should be briefly explained in the abstract, since it is not to be assumed that the term is known to every reader of the Journal.

 

• The sample size should be specified

 

• L. 43 ff. Please briefly explain why you have chosen this method and what advantages can be expected over other common methods.

 

• L. 64: Please provide sources that give the interested reader an insight into the topic.

 

• L. 64 ff. Please provide more information on the construct „Mental fatigue“. How is the state defined and what are the effects on humans.

 

• L. 85 You write: „Deep learning of complex models requires a lot of 85 training time“. That is certainly true. But how are you supposed to achieve that? An outlook on your study and how they methodically achieved sufficient training time would be useful at this point.

 

• L. 97 „The task time period method is not accurate.“ Please support with a source.

 

• L. 107 ff. I would bring this section forward to where mental fatigue is first mentioned and also elaborate more.

 

• L. 127 Please specify the sampling rate of the camera.

 

• L. 134 ff. How well could the parameters be detected? Was the error rate quantified? Please explain.

 

• Table 3: You provide a subjective assessment, but do not address it in the text. Please also write in the text why and you use subjective self-assessments in your study.

 

• L. 216 ff. Your sample consists of students. The sample size is also small. Are your results transferable to with it to the target group? Does it generate enough training data? How do the students know how to operate the simulator? Are students not overwhelmed with such an activity? Please comment on this in detail, this is a critical point.

 

• L. 261 ff.   Correlations should be statistically proven. Please calculate and state the corresponding characteristic values (even if the sample size is small as in your case).

 

• L. 317 ff. Also, please give an outlook on a future research. What do you plan to do yourself based on the results obtained and how could other research groups contribute to further research based on your results. 
•  

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper studies different methods to detect the fatigue level of humans operating the control room of a nuclear power plant. The latter environment, considered as the focus of the study, justifies the use of eyes blink rate and mouse velocity as the objective quantities allowing to evaluate the fatigue level. Indeed, the expected duties of the operator in the power plant control room are mainly related to the analysis of graphs, data and messages on multiple monitors.

The overall description of the proposed methods seems reasonable, but there are also some issues to be addressed:

- several remarks about the results in Section 3.2 are too short, if not bad written, particularly the paragraph just after Table 6.

- the best performances should be highlighted somehow in all the tables, to improve readability.

- the fact that only volunteer students have been involved in the study is actually a relevant limitation, as also acknowledged by the authors. Though I understand that enrolling real operators of a nuclear power plant may be challenging, the authors may add some remarks about the different level of skillness that could be expected from such technicians.

Author Response

Please see the attachment.

Author Response File: Author Response.docx