Design and Testing of Two Haptic Devices Based on Reconfigurable 2R Joints

Round 1

Reviewer 1 Report

The manuscript is written in a good way and readers can easily understand work. Abstract gives a very clear and concise explanation of this work. It is of scholarly nature where authors have briefly discussed two control modes for haptic device operation, selection principles, mechanical designs and statistical analysis. Conclusion is well written where authors have clearly explained their contributions and four possible future research directions. I have a few suggestions to improve quality of this work

1. In abstract, discussion on (remove) of the results is provided
2. Line 27, 3D. Define abbreviation at its first appearance.
3. Line 57, In this (these) haptic devices
4. Line 57-60, Authors have mistakenly written Section 0 for all Sections. Rewrite correct section numbers.
5. Authors should add more discussion in introduction section.
6. Line 67, authors have written FEM? Define all abbreviations at first place to make it easy and understandable for readers.
7. Line 70, we indicate the main features and analysis of those previous works. Authors should cite relevant works here.
8. Line 136, seven question (questions)
9. Author must properly define or explain variables in equations such as equ. 1,2,3 to make it more clear. 
10. Line 248, Authors should write ABS material with full form.
11. In Section 6.2, authors have clearly mentioned performance of velocity control mode is better as compared to position control mode. So any other suggestions to improve position control mode except joint manufacturing?
12. In Figure 6 (b). What is 65,105 and 136? Indicate clearly.
13. Line 275-278. Authors did not explain complete role of both sensors. Authors can add 1-2 more lines about it.
14. In Section 6.3. Authors should add significant discussion to highlight possible ways to tackle the crucial problem of clearances.
15. Authors did not mention the unit for velocity in Table 1.
16. I suggest that authors should add atleast two more relevant references from year 2021.

Author Response

First of all, we appreciate the comments and suggestions given by the reviewers in relation to the work explained in the paper, and the time they have spent studying it. We have revised the paper accordingly. Subsequently, the different issues addressed by the reviewers are answered. All the changes in this revised version are tracked with the corresponding tool.

 

1. In abstract, discussion on (remove) of the results is provided

Authors:

First, we want to thank you for your comments. The abstract has been modified according to the suggestions of the reviewer.

 

2. Line 27, 3D. Define abbreviation at its first appearance.

Authors:

The abbreviation has been defined.

 

3. Line 57, In this (these) haptic devices

Authors:

The text has been modified according to the suggestions of the reviewer.

 

4. Line 57-60, Authors have mistakenly written Section 0 for all Sections. Rewrite correct section numbers.

Authors:

This was related to converting the manuscript into PDF. Now, the section numeration has been corrected.

 

5. Authors should add more discussion in introduction section.

Authors:

The discussion on the most beneficial architecture for the haptic device has been added to Introduction section (line 49 -53).

 

6. Line 67, authors have written FEM? Define all abbreviations at first place to make it easy and understandable for readers.

Authors:

The abbreviation has been defined.

 

7. Line 70, we indicate the main features and analysis of those previous works. Authors should cite relevant works here.

Authors:

The required citations, which are [3, 28, 29] are placed in that paragraph, previous to the sentence in line 70.

 

8. Line 136, seven question (questions)

Authors:

The typo has been corrected.

 

9. Author must properly define or explain variables in equations such as equ. 1,2,3 to make it more clear.

Authors:

Authors prefer to avoid duplicating our previous work, where the detailed information on the variables and the equations can be found. A brief explanation of the variables has been added to the text. More information is available in [28].

 

10. Line 248, Authors should write ABS material with full form.

Authors:

The abbreviation has been defined.

 

11. In Section 6.2, authors have clearly mentioned performance of velocity control mode is better as compared to position control mode. So any other suggestions to improve position control mode except joint manufacturing?

Authors:

Unfortunately, the position control mode’s main disadvantage is the sensitivity to the handle position. For the human operator the difference in the inclination angle of 0.5-1 degree might not be noticeable, however, the system will respond to this small change. It could be possible to enlarge the length of the handle to make this difference in angle more noticeable for the operator; however, it will lead to increase of the overall dimensions of the haptic device and extend the range of motion of the operator’s hand to reach the positions of the workspace. The explanation of this problem has been added to the text (line 404 - 409).

 

12. In Figure 6 (b). What is 65,105 and 136? Indicate clearly.

Authors:

The description of the mentioned distances has been added to the text.

 

13. Line 275-278. Authors did not explain complete role of both sensors. Authors can add 1-2 more lines about it.

Authors:

The corresponding description has been added to the text.

 

14. In Section 6.3. Authors should add significant discussion to highlight possible ways to tackle the crucial problem of clearances.

Authors:

Yes, the reviewer is right; we are currently working on this topic. We have developed the methods of the clearance compensations for both joints in active applications. For the active 2R spherical mechanism-based joint clearance compensation is based on the difference in the commands of the actuators. In the differential gear-based joint it can be implemented by the pretension in the bevel gear, as in the previous case, by adjusting the actuators that are connected to the blue shafts of the joint (see Figure 1b). The methods of clearance compensation suitable for passive applications are still an open issue and we will approach this topic in the future work.

 

15. Authors did not mention the unit for velocity in Table 1.

Authors:

In Table 1, the velocity column is related to the name of the proper control mode, not the value, thus, there is no need in any units.

 

16. I suggest that authors should add at least two more relevant references from year 2021.

Authors:

The reference [1] has been replaced; the reference [30] has been added to the text.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript introduces two implementations of haptic feedback devices and describes their experimental evaluation in a set of user studies.

1. General considerations:

L198 “The number of cases, when the center of the end-effector goes out of the circle is counted.” – according to this description, I would assume, that if the participant could never keep up with the target circle (for example, the end-effector was always behind the projected target), you would give only a single “penalization point” to him/her, since the penalization condition would occur only once. This specific situation is unlikely, however I can imagine that some participants were unable to keep up with the goal for longer periods of time during the test. How did you penalize such situations? How could these situations affect the obtained results?

L189 – the inherent flexibility of the 2-RFR manipulator seems to make it quite difficult to control for the participants. Why did you opt for a real 2D mechanism instead of a simple 2D simulation on a screen as an environment for your experiment? Especially since a simulated environment would provide you with ways of accurately measure and quantify the end-effector deviation from the target trajectory. Could the mechanical properties of the 2-RFR manipulator affect the participants' scores?

Figure 5. presents “torque generating bushings” – a very interesting solution, which I personally have not encountered before. If this is a known solution, I would recommend adding a generic name or link to the author (if applicable). I am assuming that you need to know the relative position of the “inner and outer drums” to control the “generated” torque, however you only have a single encoder per axis. Does this mean, that you assume the outer ring position from the stepper motor? Or do you have torque sensors in the handle (L276 is not clear, see further in comments)? How did you determine the appropriate force feedback values? What is the approximate response time of the implemented force feedback system and how does this affect the “controllability” of the system (from the point of view of the participants)?

L303: How did you mitigate the “practice/training effect” that arises for the participants who have participated in both experimental setups? See: https://dictionary.apa.org/practice-effect

The "Future Work" paragraph only describes possible improvements to the existing system and experimental setup, but additional information is expected on the direction of future research and development. Are you planning to use the insights from these experiments to implement a completely new system or to apply an existing system to real-world applications?

Appendix - Table 4, Table 5: in general, tabular data is more difficult to understand than graphs, I highly recommend replacing/supplementing them with graphs (means and standard errors or standard deviations). See the following examples:

• Y. Gadre, E. Rosen, G. Chien, E. Phillips, S. Tellex and G. Konidaris, "End-User Robot Programming Using Mixed Reality," 2019 International Conference on Robotics and Automation (ICRA), 2019, pp. 2707-2713, doi: 10.1109/ICRA.2019.8793988.
• Hietanen, Antti & Pieters, Roel & Lanz, Minna & Latokartano, Jyrki & Kämäräinen, Joni-Kristian. (2020). AR-based interaction for human-robot collaborative manufacturing. Robotics and Computer-Integrated Manufacturing. 63. 101891. 10.1016/j.rcim.2019.101891.

The main contribution of the paper is rather unclear, please explain in more detail the main novelty of the paper. What are the advantages of the implemented haptic feedback systems in comparison with the state of the art? What are the possibilities of the practical application of the obtained results?

 

2. Minor:

a) Introduction:

L26: “some type of feedback” – word “some” sound rather vague in this context, please provide a more specific definition (see the definition of haptic feedback) or change the sentence.

L33-35 “… it can be said that the universal and spherical joints are ones of the most widely used in industry” – again rather imprecise and contradictory sentence. Spherical joints are definitely not the most commonly used since most mechanisms utilize simple 1D revolute joints.

L35 “They are utilized in serial robots (in the base and wrists) in delta robots and Stewart-Gough platforms” – from the point of view of implementation, a typical serial robot has 6 revolute joints.

b) Sections 2-5

Due to currently used formatting, the equations are hard to read.

L110-113 “where the mechanical energy is accumulated in the flexible bars and can be released in an uncontrollable way.” – can you specify in more detail what situation is described in this sentence?

L185 “the deviation will be equal 6 cm” – do you imply some specific deviation? The sentence is unclear. 

Figures 4-7: I would recommend adding short descriptions of individual items directly in the caption of the corresponding figures (for example: Figure 4 items 1-7).

L226 “The critical value is taken from special tables, depending on the sample size and the value of *significance-level-alfa*). – it is appropriate to also mention the exact critical value used during the analysis.

L270 – do the "strips" prevent rotation by inducing friction between them and the bushing?

L276 – you mentioned a “sensor”, but you have not specified its type. Is it a torque sensor?

L298 “Here the nodes are depicted as rectangles and the topics as arrows” – I recommend placing this description into the figure caption.

Lines 309-317 provide a description of the experiment/setup rather than the result; thus, it should be located accordingly in the text.

Why are the results of statistical analyses related to the tasks 1,2,3,4 are separated to multiple tables (Tables 1,2,3) ?

 

3. Proofreading:

The article is intelligible, however, extensive proofreading is required for better readability. Please pay attention to the structure of the sentences and their formal style, as some of the sentences in the article do not sound like a formal writing. To make it more suitable for the journal, I would advise authors to rephrase the sentences, for example L218 "bigger statistical power" - "higher statistical power”.

Here are just a few examples of errors that need to be fixed:

L46 “changing the direction an axis of one of the DOF” – “of an axis”.

Line 57-60: problem with cross-references (see “Section 0” all over the paragraph).

Line 91, 92 “*Alfa*=25” degree sign is missing.

L113 “As an exact moment of such release are hard…” – “is”.

L170-172 “he/she moves the end-effector along it until the border will be reached, resulting the end effector to stop.” – it is better to change to: “the operator moves the end effector along it until the boundary is reached, causing the end effector to stop.”

L358 “was chosen as a better one” -> “preferred by the participants”

L360 “ The majority of written comments have idea about simpler control in the velocity mode…” – “Most of the written commentaries consider the velocity mode to be more convenient…”

L379 “approaches to the border” – “approaches the border”.

L401-405 – you are mentioning the “2R spherical mechanism-based joint”, but providing a reference to an image Figure 1b, which depicts the “differential gear based joint”.

The real number of errors is quite high, and all of them cannot be written down here.

Author Response

First of all, we appreciate the comments and suggestions given by the reviewers in relation to the work explained in the paper, and the time they have spent studying it. We have revised the paper accordingly. Subsequently, the different issues addressed by the reviewers are answered. All the changes in this revised version are tracked with the corresponding tool.

 

1. General considerations:

L198 “The number of cases, when the center of the end-effector goes out of the circle is counted.” – according to this description, I would assume, that if the participant could never keep up with the target circle (for example, the end-effector was always behind the projected target), you would give only a single “penalization point” to him/her, since the penalization condition would occur only once. This specific situation is unlikely, however I can imagine that some participants were unable to keep up with the goal for longer periods of time during the test. How did you penalize such situations? How could these situations affect the obtained results?

Authors:

We have done pilot testing and we have seen that this possibility has not arisen. The tests appeared to be simple enough for the operators to keep up with the moving object most of the time. If the operator would never keep in the circle, the result of the test would have been excluded from the experimental data.

 

L189 – the inherent flexibility of the 2-RFR manipulator seems to make it quite difficult to control for the participants. Why did you opt for a real 2D mechanism instead of a simple 2D simulation on a screen as an environment for your experiment? Especially since a simulated environment would provide you with ways of accurately measure and quantify the end-effector deviation from the target trajectory. Could the mechanical properties of the 2-RFR manipulator affect the participants' scores?

Authors:

This haptic device has been developed in the scope of a wider research project focused on the kinematic study of ultraflexible parallel manipulator, so the main purpose is to test the validity of the haptic device in this scope. This has been clarified in the paper. Due to the working principles of the control system, the mechanical properties of the 2-RFR ultraflexible manipulator cannot affect the participant’s score.

 

Figure 5. presents “torque generating bushings” – a very interesting solution, which I personally have not encountered before. If this is a known solution, I would recommend adding a generic name or link to the author (if applicable). I am assuming that you need to know the relative position of the “inner and outer drums” to control the “generated” torque, however you only have a single encoder per axis. Does this mean, that you assume the outer ring position from the stepper motor? Or do you have torque sensors in the handle (L276 is not clear, see further in comments)? How did you determine the appropriate force feedback values? What is the approximate response time of the implemented force feedback system and how does this affect the “controllability” of the system (from the point of view of the participants)?

Authors:

To the best of the authors’ knowledge, there existed nothing similar to torque-generating bushing; the design shown in the paper was an evolution of a previous bushing design of the authors. There is no need to know the exact difference in the position of the drums, as the feedback in the system utilizes the values of the force from the force sensors in the handle. The torque is translated into force by knowing the dimensions of the mechanism and the forces that act in the handle. We have determined the force feedback values empirically with a preliminary testing by a separate group of participants. The response time depends on several factors. The frequency of sensor surveying is 10 Hz; however, the most crucial factor is the time of execution of the command by the stepper motors of the 2-RFR manipulator. The command execution time depends also on the control mode, as the position control mode allows motion with the higher speed, but longer time of execution.

 

L303: How did you mitigate the “practice/training effect” that arises for the participants who have participated in both experimental setups? See: https://dictionary.apa.org/practice-effect

Authors:

There were no particular measures taken to prevent the effect of training, however, an additional statistical test was carried out in order to compare the results of these 9 participants with the results of the other 10, who were testing only the differential gear-based joint. This test did not show a statistically significant difference between these two groups. It can be caused by a significant amount of time between these two tests, as due to the necessity of rebuilding the experimental setup and COVID-19 restrictions there could be up to two weeks between the tests.

 

The "Future Work" paragraph only describes possible improvements to the existing system and experimental setup, but additional information is expected on the direction of future research and development. Are you planning to use the insights from these experiments to implement a completely new system or to apply an existing system to real-world applications?

Authors:

The aim is to carry out a research work to analyze the impact of the ultraflexible manipulator on a human being during a collaborative operation. The study will involve the assessment of controlling performance of the ultraflexible mechanism with the haptic device for this kind of collaborative tasks. The corresponding explanation has been added to the future work subsection.

 

Appendix - Table 4, Table 5: in general, tabular data is more difficult to understand than graphs, I highly recommend replacing/supplementing them with graphs (means and standard errors or standard deviations). See the following examples:

• Y. Gadre, E. Rosen, G. Chien, E. Phillips, S. Tellex and G. Konidaris, "End-User Robot Programming Using Mixed Reality," 2019 International Conference on Robotics and Automation (ICRA), 2019, pp. 2707-2713, doi: 10.1109/ICRA.2019.8793988.
• Hietanen, Antti & Pieters, Roel & Lanz, Minna & Latokartano, Jyrki & Kämäräinen, Joni-Kristian. (2020). AR-based interaction for human-robot collaborative manufacturing. Robotics and Computer-Integrated Manufacturing. 63. 101891. 10.1016/j.rcim.2019.101891.

Authors:

The authors agree with the reviewer that the graphical representation of the information is easier to perceive. However, we had to maintain the equilibrium between the length of the paper and the information that has to be presented. We have added Figures 9 and 10 to Appendix A as an example.

 

The main contribution of the paper is rather unclear, please explain in more detail the main novelty of the paper. What are the advantages of the implemented haptic feedback systems in comparison with the state of the art? What are the possibilities of the practical application of the obtained results?

Authors:

The main contribution and the advantages of the proposed haptic devices have been highlighted in the introduction. The practical application has been clarified in the future work (line 480-482)

 

2. Minor:
3. a) Introduction:

L26: “some type of feedback” – word “some” sound rather vague in this context, please provide a more specific definition (see the definition of haptic feedback) or change the sentence.

Authors:

The definition of the haptic device has been modified.

 

L33-35 “… it can be said that the universal and spherical joints are ones of the most widely used in industry” – again rather imprecise and contradictory sentence. Spherical joints are definitely not the most commonly used since most mechanisms utilize simple 1D revolute joints.

Authors:

We have clarified the sentence as we are mostly focused on parallel manipulators and in this field spherical and universal joints are ones of the most commonly used. These types of joints are kinematically close to the ones considered in this paper.

 

L35 “They are utilized in serial robots (in the base and wrists) in delta robots and Stewart-Gough platforms” – from the point of view of implementation, a typical serial robot has 6 revolute joints.

Authors:

Yes, the reviewer is right; we have removed the reference to the serial robots.

 

1. b) Sections 2-5

Due to currently used formatting, the equations are hard to read.

Authors:

The authors have followed the format of the template of the journal.

 

L110-113 “where the mechanical energy is accumulated in the flexible bars and can be released in an uncontrollable way.” – can you specify in more detail what situation is described in this sentence?

Authors:

A brief explanation of this phenomenon aside with the reference where more information can be found, have been added to the text.

 

L185 “the deviation will be equal 6 cm” – do you imply some specific deviation? The sentence is unclear. 

Authors:

The sentence has been clarified in the text (line 188 -192).

 

Figures 4-7: I would recommend adding short descriptions of individual items directly in the caption of the corresponding figures (for example: Figure 4 items 1-7).

Authors:

The item description has been added to Figures 4-7.

 

L226 “The critical value is taken from special tables, depending on the sample size and the value of *significance-level-alfa*). – it is appropriate to also mention the exact critical value used during the analysis.

Authors:

The authors assume that it will not add more value to the paper without more detailed consideration of the statistical analysis.

 

L270 – do the "strips" prevent rotation by inducing friction between them and the bushing?

Authors:

Yes, this is the main working principle of the locking mechanism.

 

L276 – you mentioned a “sensor”, but you have not specified its type. Is it a torque sensor?

Authors:

The sensor type (tension-compression load cell) has been specified in the text.

 

L298 “Here the nodes are depicted as rectangles and the topics as arrows” – I recommend placing this description into the figure caption.

Authors:

A corresponding description has been added to the caption of Figure 8.

 

Lines 309-317 provide a description of the experiment/setup rather than the result; thus, it should be located accordingly in the text.

Authors:

Yes, the reviewer is right, the authors have renamed the section to make it clear.

 

Why are the results of statistical analyses related to the tasks 1,2,3,4 are separated to multiple tables (Tables 1,2,3) ?

Authors:

The results were presented in three tables because each table represents the result of a certain group of statistical analysis. For instance, the analysis of the force feedback system performance (Table 1) is not directly related to the joint control mode comparison (Table 2) and should be presented in a different format, which makes uniting the tables impossible. 

 

3. Proofreading:

The article is intelligible; however, extensive proofreading is required for better readability. Please pay attention to the structure of the sentences and their formal style, as some of the sentences in the article do not sound like a formal writing. To make it more suitable for the journal, I would advise authors to rephrase the sentences, for example L218 "bigger statistical power" - "higher statistical power”.

Authors:

The text has been subjected to proofreading. The correction proposed by the reviewer has been implemented in the text.

 

Here are just a few examples of errors that need to be fixed:

L46 “changing the direction an axis of one of the DOF” – “of an axis”.

Authors:

The typo has been corrected. We have also reviewed the whole text correcting some other typos.

 

Line 57-60: problem with cross-references (see “Section 0” all over the paragraph).

Authors:

The section numeration has been corrected.

 

Line 91, 92 “*Alfa*=25” degree sign is missing.

Authors:

The degree sign has been added.

 

L113 “As an exact moment of such release are hard…” – “is”.

Authors:

The typo has been corrected.

 

L170-172 “he/she moves the end-effector along it until the border will be reached, resulting the end effector to stop.” – it is better to change to: “the operator moves the end effector along it until the boundary is reached, causing the end effector to stop.”

Authors:

The text has been modified according to the suggestions of the reviewer.

 

L358 “was chosen as a better one” -> “preferred by the participants”

Authors:

The text has been modified according to the suggestions of the reviewer.

 

L360 “ The majority of written comments have idea about simpler control in the velocity mode…” – “Most of the written commentaries consider the velocity mode to be more convenient…”

Authors:

The text has been modified according to the suggestions of the reviewer.

 

L379 “approaches to the border” – “approaches the border”.

Authors:

The text has been modified according to the suggestions of the reviewer.

 

L401-405 – you are mentioning the “2R spherical mechanism-based joint”, but providing a reference to an image Figure 1b, which depicts the “differential gear based joint”.

Authors:

The text has been modified according to the suggestions of the reviewer.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The manuscript was improved, however, I would recommend integrating the Author Responses into the manuscript as some readers may have the same questions.