Neural Network Learning Algorithms for High-Precision Position Control and Drift Attenuation in Robotic Manipulators
, Adam Wolniakowski 1, Nesrine Kadri 2, Mateusz Sewiolo 1 and Lorenzo Scalera 3Round 1
Reviewer 1 Report
The authors propose to replace the default speed controller for high-precision position control and drift attenuation in robotic manipulators with a learning model (ANN).
Below are my comments to improve the presentation of the article:
1. In a very short abstract, try to highlight your contribution.
2: introduction: add a paragraph at the end as a summary of the rest of your paper;
3: Try to put a related work section after the introduction section;
4: Add a figure of your approach with all the steps. and describe each step in a subsection.
5: Results: Try to compare your results with those of other models, like machine learning.
6: Enrich the results section more. I find that these results can be enriched by others.
7/ reference: your references before 2021; try adding 2022 and 2023 jobs.
Author Response
Please see the enclosed file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
1-In the section : 3. Experimental results: You should replace figure 6 instead figure 7 ( You have in total 6 figures)
2- in the same section (3. Experimental results): more explanation are required: Why in figure6 ( a0 and b0): q5 Vs(time) LM is not matching with desired joint position and velocities??
only small errors are there
Author Response
Please see the enclosed file.
Author Response File: Author Response.pdf
Reviewer 3 Report
1. The originality of the paper is poor; in fact, the approaches are well known, and many related works can be found in the literature. The authors should clarify their innovations.
2. The authors must make a better effort at referencing the significant papers. There are several papers published in MDPI dealing with the subject that are not cited.
3. In addition, I need to see a conclusion of related work in the form of a table in terms of evaluation tools, performance metrics, datasets, advantages, and disadvantages that can be reconciled from other researchers' work to your own.
4. The authors should integrate the kinematic, dynamic modeling (position, velocity, acceleration, etc.), singularities, and workspace of the robot.
5. Why did the authors use only one trajectory in the xy plane (z constant)?
6. According to tables 2 and 3, we do not really see the advantage of the proposed controllers compared to the PID; on the contrary, the RMSE obtained with the PID is better! The authors need to give more explanation.
7. The authors should add a representation of Table 2 in the task space.
8. The authors need to talk about the dataset used in this paper. Because the accuracy of the robot heavily depends on the quality of the data.
9. What are the shortcomings (limitations) of this method? More detailed analyses and discussions on this are necessary. I would also like to see a frank discussion of the limitations, realism, and potential implementation problems with the system. Certainly, overhead is one issue to examine as an implementation issue of their approach.
10.There are many symbols and abbreviations. A list (nomenclature) should be given.
Author Response
Please see the enclosed file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Thank you for your efforts to respond to most of my comments and I wish you good luck
Reviewer 3 Report
Thanks for the point-by-point responses. All comments have been responded to. The manuscript quality is enhanced based on the revised version. I don't have any further comments. The paper can be accepted for publication.
