Design of JET Humanoid Robot with Compliant Modular Actuators for Industrial and Service Applications

Round 1

Reviewer 1 Report

The article describes the development of a humanoid robotic platform, JET.  Although the article seems promising, since the focus is purely on the design of the platform and concerns and implications of that design, I am not convinced that the article fits the scope of this journal and would benefit more from a submission to a Robotics oriented journal.

 

Furthermore, the article should include comparisons with other state-of-the-art humanoid robot designs if the authors are proposing this design as a novel one.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Congratulations for the paper. Currently the state of the art of robotics requires above all a practical application of technology. This paper contains an analysis of that practical application and also uses commercial systems for mange the robots.
The development of the paper is adequate and complete. The most interesting thing, in my opinion, is the test with the robot getting off the vehicle. It is a process that requires a large number of moving actuators and that forces the robot to go through positions that can be used in other processes on an assembly line: leg flexion, hip twist, torso tilt, arm movement , etc...
Because of its content, the practical aspect of the paper, and its development, in my opinion, it could be of interest to other research groups. For that reason I recommend its acceptance.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript presents the so-called JET humanoid robot platform, which enhances an  unspecified version of  the THORMANG platform by ROBOTIS. While this report on the JET humanoid robot platform development is nicely written and a joy to read, it does not provide any answer  to the most important question:  How does the results presented enhance the state-of-the-art in humanoid robotics?

 

The JET humanoid robot platform is compared only with the THORMANG platform and the version of the THORMANG platform is not even specified. The manuscript refers to the DARPA Robotics Competition from 2015, which was before the introduction of the  THORMANG platform version 3. What is even worse, the manuscript does not provide exact data comparing JET  to THORMANG nor does it describe the corresponding experiments in sufficient detail. Some comparison can be found in Table 2 and Figure 6, but without specifying the version of THORMANG (it should be the latest one or one with the same parameters). Section 4, which should report on the performance of the JET  platform, only recounts already published information and very vaguely describes two experiments ( stair climbing and rising from a seat). The experiments should be described in detail sufficient to  demonstrate their reliability and validity.

When comparing the achieved results to the state-of-the-art, please do not forget that the world outside of the THORMANG platform and DRC exists, too.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

1. Very interesting research on robot compliance with some other design enhancements to an existing system in a very well drafted manuscript that needs some mild revisions.
   • The proposed design is strongly expressed, while the compliance proposals are weakened by age of approach, lacking the latest published innovations.
2. Abstract is okay but is not likely to entice the readership to continue reading the rest of the manuscript.
   • Use of acronyms/abbreviations in an abstract is unlikely to attract readers not already aware of the manuscript’s content. Please consider whether the reader is helped by the use of LET, THORMANG, and ROM in the abstract.
   • Results are only presented in weak, qualitative fashion. Highest quality expression of main conclusions or interpretations is quantitative results discussed in the broadest context possible, e.g., percent performance improvement compared to a declared benchmark. “…enhancements reflected the design concepts…” is very weakly stated results compared to “…xxx percent performance improvement in hop roll angle over conventional methods was achieved….”
3. Introduction is decently done with some omitted very recent literature and some mild abuse of multi-citation without elaboration.
   • Deterministic artificial intelligence (as recently applied to DC motors, unmanned underwater vehicles, and even space systems) was omitted alternative to torque-based or impedance-based controllers. It seems D.A.I. would permit any actuator (including the SHG-17-100-2SO and H54-200-S500-R) to produce desired stiffness to achieve specified compliance to counter pelvis tilt aiding control of foot swing, but the technique is not mentioned as suggested future research.
   • Another omitted recent innovation as a competing alternative to torque-based or impedance-based controllers is optimal robot whiplash compensation that uses Pontryagin’s minimization of Hamiltonian systems to derive controls that account for interaction with robot structural dynamics to meet multiple objective functions, including compliance.
   • Please elaborate a reason for the reader to investigate each of the triple cited references [3-5].
   • Please elaborate a reason for the reader to investigate each of the quintuple cited references [9-13].
4. In section 2, please add citation permitting the reader to investigate Dynamixel Pro, SHG-17-100-2SO harmonic drive and all other named-systems required to be understood by the reader to duplicate the research.
5. Figures are very well done and enhance the manuscript considerably, especially amidst a dearth of analytical development. Consider further emphasizing the experimental nature of the paper to persuade other potential peer reviewers (experimental papers are often shunned unfairly by academic peers).
6. Tables are decently done to introduce problem formation (aiding repeatability), but quantitative results are neglected.
7. Inclusion of the appendix for acronyms would be a welcome addition to the appendix.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

While there are some changes in the manuscript and I really appreciate  the way in which the cover letter is presented, critical issues are not sufficiently addressed and the manuscript itself is still almost all “applied” and little to no “sciences”. Even after the changes, the paper still fails to present any scientific novelty (for 2021) and the experiments are not sufficiently described. This conclusion is based on the following facts:

• The newly added references on lines 34-37 do make the list of references look better but that's all. The lines 34-37 do not compare the new references to the presented results in any way and they are not mentioned in the rest of the manuscript at all.
• A comparison to the TALOS  humanoid robot is added to some parts of the manuscript. But TALOS is a commercial solution, available since 2017. The authors of the manuscript still do not confront their results with more recent RESEARCH works.
• The authors of the manuscript have already published several papers about their JET humanoid robot platform.While in this manuscript the authors provide some not yet published information about the JET platform itself, it is the evaluation of the platform, which should demonstrate its novelty and advantages with respect to other solutions. I appreciate that Figure 6  has been updated with data about THORMANG3 (2016) and TALOS (2017),  albeit data about more recent platforms will be welcomed.
  But the evaluation in Section 4 still fails to deliver. The experimental setup and procedures of the “stair climbing” and “rising from a seat” experiments should be described in detail sufficient to replicate them. For example, the  dimensions of the stair steps are not given. In the “rising from a seat” experiment the position of the handle, which the robot uses when standing up should be given, too. Exact data should be provided and not formulations like “similar height of the typical counter stool” or “a little shorter”. The results of the evaluation should also be provided in exact form. But the biggest problem is that JET  is compared only to the THORMANG of Team SNU, which according to the information provided in the manuscript is at least six years old.

 

Without addressing these issues I do not see the manuscript fit for a publication in a scientific journal.  I do believe that the authors failed to improve their manuscript sufficiently in the first round of the revision because of insufficient time. Therefore I recommend another revision.

Author Response

Please see the attachment

Author Response File: Author Response.pdf