Design and Evaluation of an Underactuated Rigid–Flexible Coupled End-Effector for Non-Destructive Apple Harvesting

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The title topic is in Design and Evaluation of an Underactuated Rigid–Flexible Coupled End-Effector. It is a very interesting topic. 

Typos: 
In line 144: "Figure 1a". But the authors used Figure 1 (a), which 1 do you choose, "(a)" or "a"? And so on. 

In line 222: "𝜔represents the virtual angular velocity."
It should be "𝜔 represents the virtual angular velocity."
Please check the other notations. And so on.

In line 226: "force.Where:"
It should be "force, where:"
Please check the other notations. And so on.

Where (in line 248) should be "where". Please check the others "where".

In line: 639: "The specific validation results are presented in Table 4."

Actually, Table 4 can't be found.
I thought in line 640: "Table 2. Verification results" should be Table 4 (after Table 3).

The Figures, Tables and the caption or Sections should not be presented in different page. Also in other Figure. For example: See on Figure 6 or Section Conclusion.

The distance between the figures and tables and previous or next paragraphs should be in consistent. For example: The previous paragraph (line 306) and Figure 5.

The authors should use the term Figure or Fig consistently.

There two equations with the same number of equation? Equation (15) and Equation (15)?
The authors should warn about that. 
It's better, the authors use LaTEX rather than the other word processor for better layout.

Question:
In line: 147, the authors mentioned that displacement accuracy ≤ 0.0001 mm. When will the authors use that accuracy? What will be effect to the experiment results?

In Equation (1) and (2); What is the differentiate between T and T. The first T represents Torque, and the second T represents transpose? Please consider about that.

In Abstract:
In line 40-42: "The average relative error between simulation and experiment was 5.89% for flexible silicone contact force"
In Experiment Results, this average relative error between simulation and experiment couldn't be find. The authors should explain how can the authors mentioned this in Abstract?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors present design of an underactuated 3 finger rigid-flexible gripper for picking apples. The rigid-flexible underactuated gripper itself is not novel, but applying structural optimization, complex multibody flexible-rigid simulation and experimental validation on a real physical prototype add value.
 
 
 
Must be improved before publication:
 
No details are given on silicone used for PPRS, MPRS, and DPRS parts of the finger. Fabrication method can significantly influence silicone properties, so please provide details on which silicon is used, curing method, shore hardness, and how is the silicone attached to the rigid phalange in experimental setup? Provide details - using what adhesive, any peeling or slipping during experiments, and explictly state that this interface is modeled as perfectly bonded  - fixed in simulations.
 
 
Figure 2. Please provide boxplots of measured load damage per apple region instead of scatter plot on Figure 2 b. Include information on sample size in the boxplot. Also, on Figure 2. a) please explain first discontinuity – labeled damage point at approximately 60 N, and give explanation of the second discontinuity at approx. 160 N in the text. Please relate these values to the actual measured damage forces for the top, middle and bottom region.
 
 
Please provide details on hardware used for experiments, along with exact MATLAB version, toolboxes used and version of toolboxes.
 
Clearly explain the motivation for selecting fitness function as f=|Fmax -Fmin|. This way, gripper is optimized solely for uniformity, but no constraints on maximum forces are imposed. Please elaborate on limitations of this fitness function.
 
Population size 5000 seems like a large number. Please justify this number and add information on runtime with this population to reaching stopping criteria in genetic algorithm for each of the ten results in Table 1.
 
In table 2 please provide units for Contact damping.
 
In table 2 – please provide details (reference or method) on how Poisson’s ratio and elastic modulus is determined for apple.
 
Equation 23 is not correct. -please add square in denominator – Rc must be squared.
 

 

Minor:
 

 
Please explain that relation given in eq. 21 holds only if silicon surface is considered planar with radius significantly larger than the radius of the apple.
 

 
In table 4.:
 
To strengthen the validation please report explicitly the number of experimental trials (i.e. 10) and the variability of the measured contact forces (mean ± stanard deviation). This would allow readers to better interpret the relation of MAPE and the variability of experimental measurements.
It would be helpful to mention in discussion that interface between silicone and rigid finger part is modelled as fixed joint, and that ths voids possible interface slip or debonding. This makes the assumptions behind the coupled rigid-flexible model more clear. 
 
 

repeated sentence:
line 404
During the gripping process, the load exerted by the end-effector on the apple must remain below the apple’s damage threshold, the load exerted by the end-effector on the apple must remain below the apple’s damage threshold

Table layouts are not uniform across paper.

Please check typos, reference inconsistences.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This paper addresses the need for non-destructive grasping in apple harvesting robot end-effectors by proposing a rigid-flexible coupled end-effector based on the principle of underactuation, which is systematically evaluated through simulation and experimentation. The research has a clear engineering application background, employs a relatively systematic methodology, and yields results with certain reference value. However, there remains room for improvement in several aspects of the paper. The specific review comments and revision suggestions are as follows:

1. The abstract is overly lengthy. It is recommended to condense it, highlighting the research objectives, methodology, key results, and conclusions.
2. The literature review in the Introduction is not sufficiently systematic, particularly regarding the research progress on the combination of "rigid-flexible coupling" and "underactuation." It is suggested to supplement recent relevant literature and more clearly articulate the innovation points of this study.
3. In Section 2.1.1, please clarify whether the selection of apple samples is representative (e.g., in terms of maturity, orchard management conditions) and whether the influence of storage conditions on mechanical properties has been considered.
4. In Section 2.3 on genetic algorithm optimization, please provide a detailed explanation for the design rationale of the fitness function, the basis for setting parameter ranges, and the scientific justification for the algorithm's termination criteria.
5. In Section 2.4 on simulation modeling, it is necessary to explain why SOLID187 and SOLID185 elements were chosen and whether mesh sensitivity analysis was conducted to verify mesh independence.
6. For the experimental platform in Section 2.5, details such as the force sensor calibration method, sampling frequency, and data filtering methods should be supplemented to ensure data reliability.
7. In Section 2.6 on damage assessment, the application of Hertz contact theory to the apple-silicone interface (non-homogeneous materials) requires caution. It is recommended to discuss its applicability and limitations.
8. In the simulation results of Section 3.1, quantitative analysis of force distribution uniformity during the grasping of apples of different sizes (e.g., coefficient of variation) should be added. The experimental sample size is small (only one sample per size for three sizes?). It is recommended to increase the number of repeated trials and perform statistical analysis.
9. The Discussion section is weak. It should delve deeper into the sources of error between simulation and experiment (e.g., material model simplification, assembly errors, sensor noise) and discuss their potential impact on practical application. There is a lack of comparative analysis with the performance of existing similar end-effectors (e.g., grasping force, damage rate, operation time), making it difficult to demonstrate the superiority of the design presented here.
10. The paper emphasizes the combination of "rigid-flexible coupling + underactuation" as an innovation, but the explanation of the mechanistic advantages of this structure in apple harvesting is insufficient. It is recommended to elaborate further in the Introduction or Discussion. Please clearly state the applicability and potential challenges of this study in real orchard environments (e.g., lighting, branch/foliage occlusion, variable fruit posture). The Conclusion section should be more concise, avoiding repetition of result content, and can include brief statements on study limitations and future prospects.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed the majority of my comments, provided justifications and explanations where applicable.   The manuscript is now more scientifically sound and suitable for publication.

As a suggestion for future work, when employing genetic algorithms, I recommend performing an initial sensitivity analysis of the algorithm parameters. This would enable a more rigorous and justified selection of GA settings, moving beyond default values, and could potentially reduce runtime while improving solution quality. In addition, logging optimization runtime should be routinely performed to support reproducibility and transparency of optimization results.

Finally, I recommend  authors to check typos once again, like missing space here: 

[22, 28, 20, 21, 2, 3]and upper, line 278 on page 9.

Author Response

We thank the reviewer for the insightful suggestions regarding our future work. We will ensure that parameter sensitivity analysis is performed and runtime, along with other detailed parameters, is logged in our future studies to enhance rigor and reproducibility.
Furthermore, we have thoroughly checked the manuscript for typing errors. The missing space pointed out by the reviewer, as well as other similar issues, have been corrected and marked in red in the revised manuscript.
Finally, we thank you for your contribution to this manuscript.

Reviewer 3 Report

Comments and Suggestions for Authors

None.

Author Response

We thank the reviewer for their review and support of our manuscript.