RREV: A Robust and Reliable End-to-End Visual Navigation

Round 1

Reviewer 1 Report

I found the authors' work very interesting. Their manuscript is easy-to-follow and, generally, well written. However, I doubt if the presented work concerns the journal's scope since the authors' method is not referred to drones but wheel vehicles.

Some minor comments:

Please use the present tense in your manuscript. E.g., in the abstract and the introduction, when someone reads, your manuscript should be in the present since this method is not old. The authors can use past tense in the related work, which is more normal.

Introduction: high-efficiecy ->highly-efficiently

face following problems -> face several problems

In general a proof-reading from an expert is needed.

Author Response

Dear reviewer,

Thank you very much for your kindly comments on our manuscript. There is no doubt that these comments are valuable and very helpful for revising and improving our manuscript. Our response to your suggestion is as following:

General Comments: I found the authors' work very interesting. Their manuscript is easy-to-follow and, generally, well written. However, I doubt if the presented work concerns the journal's scope since the authors' method is not referred to drones but wheel vehicles.

Point 1: Please use the present tense in your manuscript. E.g., in the abstract and the introduction, when someone reads, your manuscript should be in the present since this method is not old. The authors can use past tense in the related work, which is more normal.

Response 1: Thank you very much for your reminder. We apologize for this syntax problem and corrected it according to your suggestions. In addition, We submit the special issue "Explainable Deep Architectures for Saliency-Based Autonomous Vehicle Driving Monitoring" under Droners. We have learned the information and contribution requirements of the special issue, and we think that the manuscript we submitted meets the requirements of the special issue.

Thank you again for your positive and constructive comments and suggestions on our manuscript.

We hope you will find our revised manuscript acceptable for publication.

Reviewer 2 Report

In this paper, a more robust and reliable end-to-end visual navigation scheme (RREV navigation) is proposed to solve the problem of imbalance between the “junctions” and “non-junctions” samples in the road scene, and the scheme is used to predict the future waypoints of vehicles from the front view RGB images. The problems of confidence evaluation and anti-disturbance are also studied in this paper. Although numerical simulation and ground tests show the superiority of the algorithm, there are still the following problems to be solved.

1. Is it possible to eliminate the imbalance by using artificially manufactured virtual samples with a reasonable proportion of “junctions” and “non-junctions”, rather than by using the complex method designed in this paper? This relates to the practicability of this paper.

2. There is a missing period at the end of the first paragraph of Part 3.1.

3. “Because in this paper, we focus on the problem of sample imbalance and robustness of deep learning based system, we hope that the input is as simple as possible.” Although this paper explains the author's understanding of robustness, does it violate the conventional definition of robustness to use input as simple as possible? Please comment. In addition, some novel related works have been appeared recently,  e.g.“Coupled orbit-attitude dynamics and trajectory tracking control for spacecraft electromagnetic docking”; “Active Disturbance Rejection Control for Delayed Electromagnetic Docking of Spacecraft in Elliptical Orbits”. Could the authors  make some descriptions about such methods in the introduction part, which may be useful in certain sense yet from different aspects.

4. In the part of “Calculate Curve Fitting Error”. Why does the paper fit the data sequence into a quadratic polynomial? What is the basis of the fitting method? Please comment or provide relevant references.

5. Why take the maximum value of F1 as the optimal value of threshold th? Is the purpose of this choice for the safety of driving? Whether such a choice is too conservative.

6. In this paper, the physical test is used to make the results more convincing, and it is suggested to add more complex road conditions to verify the robustness and reliability of the algorithm.

Author Response

Dear reviewer,

Thank you very much for your kindly comments on our manuscript. There is no doubt that these comments are valuable and very helpful for revising and improving our manuscript. We would like to answer the questions you mentioned and give detailed account of the changes made to the original manuscript. Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper addresses a new robust navigation technique for autonomous driving. This topic is of high interest and the paper is good in general, but some recommendations can be followed by the authors in order to improve the quality of the document.

1. English language should be revised throughout the document. Minor grammar adjustments have to be done.

2. Some form aspects can be improved. For instance, there is no space between citations to references and the previous word (e.g. line 27, 40, 42, 54, 58…)

3. Section 2 belongs to the introduction section. It must be moved to the beginning and the paper can be reorganized. This will allow the authors to improve the state of the art, which is a bit short in references.

4. The introduction section needs to be expanded, with material from Section 2, and new references that allow the reader to get a better idea of why autonomous driving is important, and what other techniques have been used and why this new one that is presented moves forward the state of the art.

5. The methods section also needs re-organization to allow the reader following the proposed method. The are paragraphs that need to be rewritten.

Author Response

Dear reviewer,

Thank you very much for your kindly comments on our manuscript. There is no doubt that these comments are valuable and very helpful for revising and improving our manuscript. We would like to answer the questions you mentioned and give detailed account of the changes made to the original manuscript. Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

The authors have made the corrections.