Intelligent Control Strategy for Robotic Manta via CPG and Deep Reinforcement Learning
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis paper presents a bionic manta ray controller using joint CPG-DDPG control. This approach combines CPG motion control strategy and DDPG reinforcement learning path planning strategy. The authors explain the principles of the CPG motion control strategy and the DDPG reinforcement learning strategy from a theoretical point of view, and conduct experiments in Webots simulation and a real environment, respectively, to demonstrate the superior performance of the CPG-DDPG strategy over the CPG-PID strategy and the SIN-DDPG strategy.
However, I found that the experimental portion of your paper does not make a very strong case for the superiority of your method relative to the methods in common use today. Specifically as follows:
1) Does your approach have the same clear advantages over other strategies now commonly used for bionic manta rays?
2) Is it possible to collect process data and plot it during physical experiments to give a stronger picture of the actual results of your program?
I encourage the authors to carefully consider our feedback and revise their paper accordingly. I wish the authors the best of luck with their future research.
Comments on the Quality of English LanguageNone
Author Response
Dear Reviewer:
Thanks for your comments concerning our manuscript entitled "Intelligent control strategy for robotic manta via CPG and deep reinforcement learning" (ID: drones-3039220). Those comments are valuable and helpful for revising and improving our paper and have important guiding significance for our research. We have studied the comments carefully and have made corrections, which we hope to meet with approval. The responses have been compiled into the attached PDF document.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe paper proposes a DDPG (Deep Deterministic Policy Gradient) control strategy based on CPG (Central Pattern Generators).
The CPG strategy simulates the swimming behavior of a robotic manta, and the DDPG algorithm implements a controller that adaptively changes the CPG's control parameters and allows smooth transitions between different swimming modes. The Hopf oscillatoris used to construct the motion control network.
The paper makes a modification of the CPG model, solving the problem of an excessive number of input parameters.
The paper is relevant because considers effective approaches to building control of promising bionic underwater robots. The paper is of significant significance, suggesting further development towards the construction of management of functionally complex systems.
The effectiveness of the method, quantitative and qualitative, is confirmed by the experiments presented.
The conclusion summarizes the work performed in detail. References are given correctly.
All mathematic formulas are given correctly. The figures and tables accurately represent the results.
Author Response
Dear Reviewer:
Thanks for your comments concerning our manuscript entitled "Intelligent control strategy for robotic manta via CPG and deep reinforcement learning" (ID: drones-3039220).
Reviewer 3 Report
Comments and Suggestions for Authors
This draft proposes an innovative way to control a manta-inspired robot, using a combination of Central Pattern Generators (CPG) and Deep Deterministic Policy Gradient (DDPG). The main contribution of the draft includes the development of a robotic platform with versatility, the introduction of a combined algorithm for better control strategy, and clear validation using both simulation and experimental results. The draft is overall of good quality. I have attached a list of comments and suggestions, regarding scientific content and presentation. After addressing these comments and moderate changes, the draft should be qualified to be published.
Comments for author File: Comments.pdf
Author Response
Dear Reviewer:
Thanks for your comments concerning our manuscript entitled "Intelligent control strategy for robotic manta via CPG and deep reinforcement learning" (ID: drones-3039220). Those comments are valuable and helpful for revising and improving our paper and have important guiding significance for our research. We have studied the comments carefully and have made corrections, which we hope to meet with approval. The responses have been compiled into the attached PDF document.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThis paper deals with an intelligent control strategy for robotic manta via CPG and deep reinforcement learning. I think this paper is well organized and also well described the object and main contents, This paper also well demonstrated for proposed method. However, I hope authors add the summarize for previous work by using table to represent their advantage and shortfall to easily understand by reader.
Comments on the Quality of English LanguagePlease check the English grammar.
Author Response
Dear Reviewer:
Thanks for your comments concerning our manuscript entitled "Intelligent control strategy for robotic manta via CPG and deep reinforcement learning" (ID: drones-3039220). Those comments are valuable and helpful for revising and improving our paper and have important guiding significance for our research. We have studied the comments carefully and have made corrections, which we hope to meet with approval. The responses have been compiled into the attached PDF document.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsI have no more comments.
Reviewer 3 Report
Comments and Suggestions for AuthorsI appreciate the additional work accomplished by the authors. They have addressed all the issues. The quality of the draft has been greatly improved and ready for publication.
Reviewer 4 Report
Comments and Suggestions for AuthorsI would like to decide as an "accept"
Comments on the Quality of English LanguagePlease check English grammar.