Robust Distributed Rendezvous Using Multiple Robots with Variable Range Radars
Round 1
Reviewer 1 Report (Previous Reviewer 1)
Although the authors have made a few revisions to their original manuscript, they failed to resolve the core issues. Due to the poor presentation and the weak evaluation, I would suggest a rejection. Please see the following comments and consider how to improve your manuscript.
1. The contribution is not clear. In the introduction, the authors should try to identify the research gap and emphasize the unique contribution of this paper. References should be provided to support your claim.
Moreover, the presentation looks like a technical report, not a scientific manuscript. The authors should pay more attention to the presentation of the manuscript.
2. In addition to the simulation, real robot experiments should be provided. A simple simulation cannot support your contribution.
3. More state-of-the-art algorithms should be introduced to evaluate your proposed method in simulation and real robot experiments.
Author Response
The response to Reviewer 1 is attached.
Thank you very much.
Author Response File: 
Author Response.pdf
Reviewer 2 Report (New Reviewer)
The authors propose a distributed rendezvous control mechanism in multi-robot systems assuming 3D environments in order to recover/maintain network connectivity. The main contribution is that the proposed approach adaptively controls the robot’s radar footprint by increasing the transmission power level by adjusting the amplifier in the transmitter. There are some modifications that the authors need to address before the paper is ready for publication:
1- It would be better if the authors discuss in the "Introduction" section some applications where the proposed approach can be used.
2- The "Introduction" section needs modification as there is repletion for presenting some concepts or ideas. For example, in the Introduction, the idea of the statement " We consider a practical scenario in which a radar sensor contains measurement noise, and the environmental disturbance generates process noise in a robot’s maneuvering" was mentioned before in the statement "Our article considers a practical scenario in which a radar sensor contains measurement noise, and the environmental disturbance generates process noise in a robot’s maneuvering ".
3- . The "Literature review" section might be strengthened by listing each reference's advantages and disadvantages.
4- The "Simulation Results" section should be strengthened by including more findings. For a journal publication, the results that are shown are insufficient. The authors may, for instance, research the effects of altering the size of the area and the number of robots.
Author Response
The response to Reviewer 2 is attached.
Thank you very much.
Author Response File: Author Response.pdf
Reviewer 3 Report (New Reviewer)
In the paper , it is not explained how to determine the constant of β in (9)。 Please give some supplement.
Author Response
The response to Reviewer 3 is attached.
Thank you very much.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report (Previous Reviewer 1)
The authors have clarified some comments.
Reviewer 2 Report (New Reviewer)
The authors have addressed most of the comments.
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
This paper presents a distributed rendezvous algorithm in cluttered 3D environments. Lots of simulations have been conducted to demonstrate the effectiveness of the proposed rendezvous approach.
The proposed idea sounds interesting. However, the following comments should be further clarified.
1. The contribution is not clear. In the introduction, the authors should try to explain their research motivation, the original idea, and emphasize the unique contribution of this paper. Where is the research gap? What is your contribution?
2. In Section 4, the authors present MATLAB simulation results to demonstrate the effectiveness of the proposed distributed rendezvous algorithm. However, merely simulation is far from enough. It is suggested to evaluate the proposed method on real robots.
Moreover, more state-of-the-art algorithms should be introduced to compare with the proposed method in both simulation and real robot environments.
3. The conclusion part should be further enhanced to summarize the main findings of the manuscript. The limitations of this study and directions for future work should be included.
Author Response
Thank you very much for your valuable comments. The response to Reviewer 1 is attached.
Author Response File: Author Response.pdf
Reviewer 2 Report
Let us reassert some points of the paper
- robots can only detect one another if there is an unobstructed line of sight between them,
- increasing the footprint of the radar cannot cause a detection in case of obstructed line of sight,
- the circumcenter algorithm (lines 191-194) calculates a "circumcenter" point for a robot based on the set of its neighbors,
- the velocities of a robot (equation 7) are calculated based on positions of other robots in a given iteration (equation 6),
- the proof of convergence given in [16] was done for an obstacle-less environment,
- movement is subject to disturbances.
There are several points that need to be addressed
- when there are no neighbours within sensor range the circumcenter algorithm fails to compute; in a special case, when all robots are isolated (either initially or throughout the execution of the algorithm) the robots fail to move at all;
- when obstacles are concerned, there is no guarantee that moving according to (7) will maintain connectivity if obstacles are involved - the "moves" for each robot are calculated independently so it is entirely possible that relative to an old (in previous iteration) position of a robot A, the new position (in the update iteration) of robot B will maintain connectivity (and vice versa) while new positions of robots A and B will not (basically, just one robot moving will not break the connectivity, but both moving will); this can isolate some/all of the robots (see previous point);
- the paper refers to [16] for the proof of convergence, however, it does not consider the modification of the circumcenter algorithm due to collision avoidance (equation 11); to claim convergence, the proof has to be redone;
- the convergence of the algorithm completely ignores the issue of disturbances which can lead to infeasibilities; for example, given big enough disturbances the robots can all move out of detection range and/or become obstructed from one another - the algorithms then fails.
Those are the major points that put into doubt several claims of the paper (especially the claim regarding the convergence of the algorithm). They need to be addressed before the paper can be put into print.
Additional notes:
- notation for a directed edge in graph A shouldn't be a set but a tuple (pair) instead,
- there are some very minor editorial mistakes (for example, missing space in line 28).
Author Response
Thank you very much for your comments. The response to Reviewer 2 is attached.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors failed to improve upon the failings of the circumcenter generation algorithm (lines 198-202).
Perhaps to better explain the issue let us consider a simple example.
Let there be two robots and a single obstacle between them (initially or otherwise). According to lines 157-160 the robots do not detect one another. Since neither robot has any detectable neighbors the projection (line 199) cannot happen. The computation of circumcenters (line 200) cannot happen. The movement (lines 201-202) cannot happen. The algorithm fails to proceed.
This failure will happen every time a robot becomes disconnected from all of its neighbors which can happen if the movement algorithm is given by (7).
Since the algorithm can simply fail to work, I cannot, in good conscience, agree to publishing those finding until the shortcomings of the algorithm are eliminated. Since they have not been fixed (and perhaps it is not trivial to do so), I propose that the paper be rejected at this point in time.
Author Response
The response to Reviwer 2 is attached. Thank you very much.
Author Response File: Author Response.pdf
