Driver Take-Over Behaviour Study Based on Gaze Focalization and Vehicle Data in CARLA Simulator

Round 1

Reviewer 1 Report

This paper presents a study of the driver 5 behaviour in the transition between autonomous/manual mode using CARLA simulator. Following are my reservations about the paper.

1. The English language of this paper is not up to the mark.

2. It seems that the authors have extracted this paper from their funding proposal as they keep referring to the proposal.  

3. There should be a figure demonstrating the actual experimental setup. Where are the cameras placed, where the participant will sit, where will be interactive screen etc?

4. An explanation is required on how three screen demonstrations will map to real-world driving.

5. The results section is flooded with graphs with little explanation to it. 

6. No explanation for Figure 2 is given and the Figure 2 itself is quite confusing. 

Author Response

Please kindly see the attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

The issue of control transfer in conditionally autonomous and autonomous vehicles is extremely important. Such testing is done primarily on simulators for safety reasons. Eye-tracking systems are also used. They are not always mounted or worn on the driver. They can also be on the dashboard. Next, research is conducted on experimental tracks. In reality, all sorts of scenarios can happen, and it is difficult to predict them in advance.

Some comments:

1. The statement that driver behavior will be studied is very general and the conclusions obtained are rather predictable. It would be appropriate to formulate a specific research objective or hypothesis and demonstrate its truth or not.

2. At the L3 level, the driver must touch the steering wheel from time to time, only from L4 and above there is no obligation to do so. Thus, information about the need to take control can also be given by means of steering wheel vibration or, for example, seat vibration. Studies on the selection of modality of information about the need to take control have been conducted in many research centers. Thus, the sentence that: "Haptic modalities on the steering wheel can not fit these experiments because the driver is not in contact with the wheel" is not true.

3 The nomenclature for autonomous vehicles uses the term "Out-of-the-loop" rather than "Out-of-the-screen." I understand that in this case, what is meant is that he was not watching the screen.

4. A large number of diagrams and not always in the right place makes it difficult to analyze the problem. E.g. Figure 9 should be before Tables 1, 2.

5. The question is whether not to present only selected graphs and describe the rest of the results, for example, in a table. 

6. "Their mean age was 33.5, with a standard deviation of 11.9 years. They had a mean driving license of 15.3 with a standard deviation of 12.1 years."  The study group is not homogenous and is characterized by a wide spread in age as well as experience. Some studies clearly show differences in driver behavior depending on age and experience, which was not considered in this case.

7 Another question: were the drivers aware of the scenarios before the study? Did they know what the purpose of the study was?

8. Another doubt concerns simulator sickness. Were drivers evaluated from this point of view before or after the test drive? Such disorders can also affect the results.

9. In the statement: And the awareness was higher when the drivers were monitoring the driving. - there is probably nothing surprising. More interesting is what kind of non-driving activities distract a driver to such an extent that he or she is unable to react properly in a timely manner.

In conclusion, research on the transfer of control is very important and certainly interesting. In this case, the aim or hypothesis of the research must be clearly formulated and the description of the results organized. Excess information is not conducive to their analysis. It is worth trying to present them clearly, which will undoubtedly improve the reception of the presented content.

Author Response

Please kindly see the attached file.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have addressed all my concerns and I have no further question. 

Reviewer 2 Report

• Thank you for the comprehensive answers.
•