Dynamic Human–Robot Collision Risk Based on Octree Representation
, Giorgos Paraskevopoulos 1, George Michalakis 1, Isidoros Michalellis 1, Evangelia I. Zacharaki 1, Panagiotis Koustoumpardis 2 and Konstantinos Moustakas 1
Round 1
Reviewer 1 Report
Paper about collision risk assessment.
You need to provide some information about:
1. what kind of computing resources are needed for your solution (what did you use)?
2. how fast is the calculation of the dynamic map (you say "In AR, given the movements of the robotic arm, our process can successfully visualize in real time the dynamic safety zones in order to inform a user close to the working space of the robotic arm." but one would need some numbers, how fast can you compute the map on your hardware (see point 1).
3. how much memory is needed for your computation depending on the resolution?
4. You mention the connection to the robotic arm controller but this doesn't seem to be explained in the paper. How is your 3D robot model controlled? Do you run HIL or SIL simulations for the visualization prediction?
Minor issues:
Line 117: recent VR -> Recent VR
Line 138: TRR acronym?
Line 254: in fast -> in fact?
Line 286: cubes( -> cubes (
Line 352: ArUco acronym?
Author Response
Dear Reviewer,
We would like to express our sincere gratitude for your time, effort, and invaluable feedback on our research paper titled "Dynamic Human-Robot collision risk based on octree
representation." Your insights will contribute to the enhancement of our work. Here are our responses to the points you highlighted.
- what kind of computing resources are needed for your solution (what did you use)?
In terms of hardware, the experiments in Unity were run in a computer with 48 GB RAM, an NVIDIA RTX 3090 graphics card and an AMD Ryzen 9 3900x processor with 12 cores.
- A sentence was added in line 364.
- how fast is the calculation of the dynamic map (you say "In AR, given the movements of the robotic arm, our process can successfully visualize in real time the dynamic safety zones in order to inform a user close to the working space of the robotic arm." but one would need some numbers, how fast can you compute the map on your hardware (see point 1).
The calculation of the dynamic map in the afore mentioned computer can be calculated in less than 20ms. Hence, we can conclude that it can be visualized in real time.
- A comment in parenthesis was added in line 362.
- how much memory is needed for your computation depending on the resolution?
We use 48 GB RAM for our experiments, but the method can function with less RAM.
- You mention the connection to the robotic arm controller but this doesn't seem to be explained in the paper. How is your 3D robot model controlled? Do you run HIL or SIL simulations for the visualization prediction?
In our simulation the 3D robot model is animated through Unity with a set of predefined movements. For a real-time experiment, even though we did not perform one, we can easily interchange this set of movements with the moves provided from the controller of the real robotic arm. The connection between the controller and the program can be established through a serial or a wireless connection.
- A paragraph was added in line 365.
Line 138: TRR acronym?
TRR stands for Twisting, Rotational, Rotational joints.
- A comment in parenthesis was added in line 186.
Line 352: ArUco acronym?
ArUco stands for Augmented Reality University of Cordoba
- A comment in parenthesis was added in line 364.
The following errors have all been fixed in the attached manuscript.
Line 117: recent VR -> Recent VR
Line 254: in fast -> in fact?
Line 286: cubes( -> cubes (
Kind regards,
Nikolaos Anatoliotakis
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The researchers have presented a Dynamic Human-Robot collision risk based on octree representation. They present a framework for rendering static/dynamic safety zones in VR/AR. There is merit to this article as it suggests a method to improve human cognitive awareness of the potential risk due to the robot dynamic behavior, and seeks to improve saftey in environments requiring collaboration between humans and robots. We suggest some minor corrections:
1. The sentence is unclear in line 22, line 256.
2. Grammatical errors in line 31, 33, and more. The overall english quality can be improved.
3. It is advisable to provide the full form before providing the abbreviated form, such as in line 138, DoF.
4. Provide references for the equations where possible.
5. If any of the figures have been taken from simulation softwares, it is best to mention the name of the software in the figure caption.
Minor editing of English language required
Author Response
Dear Reviewer,
We would like to express our sincere gratitude for your time, effort, and invaluable feedback on our research paper titled "Dynamic Human-Robot collision risk based on octree
representation." Your insights will contribute to the enhancement of our work. Here are our responses to the points you highlighted.
- The sentence is unclear in line 22, line 256.
- The apparent sentence was changed in order to be more understandable.
- Grammatical errors in line 31, 33, and more. The overall english quality can be improved.
- The sentences in lines 31 and 33 were changed in order to be more understandable.
- It is advisable to provide the full form before providing the abbreviated form, such as in line 138, DoF.
Full form of DoF is degrees of freedom.
- A comment in parenthesis was added in line 139.
- Provide references for the equations where possible.
- A reference of the research paper from Cheng named “A study of robot inverse kinematics based upon the solution of differential equations” was added.
- If any of the figures have been taken from simulation softwares, it is best to mention the name of the software in the figure caption.
The figures shown in this paper have all been taken either from Unity environment or the environment of Matlab’s Robotic Toolbox.
- A reference to Matlab and Robotics Toolbox was made in the manuscript in line 274.
The revised manuscript has been attached and all the changes from the previous version have been highlited.
Kind regards,
Nikolaos Anatoliotakis
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
A better and clearer version of this work,
